Master Thesis
It takes two to tango? The composition of joint-audits in France and
audit quality
ABSTRACT
In this thesis I study the impact that individual auditor characteristics in joint-audits have on auditees’ audit quality. I draw on hand-collected data from a sample of listed French companies to test my expectations. I find that having a combination of one Big4 one non-Big4 auditor improves audit quality. I also find that when the auditors are both males, the overall audit quality is higher. Finally, I find that tenure and age are insignificantly associated with quality. All in all, the results suggest that auditor characteristics seem to impact audit quality differently relative to single-audit settings. I attribute these effects to the highly collaborative nature of joint audits which likely determines audit quality. My thesis contributes to the evolving literature stream on joint-auditing by showing that individual characteristics may impact audit quality.
Keywords: joint-audit; audit quality, group composition, French setting
Arjan PORTEMA S2812371
a.portema@student.rug.nl
MSc Organizational & Management Control Faculty of Economics and Business
University of Groningen Supervisor: dr. V.A. PORUMB
2
Table of Contents
INTRODUCTION ... 3
THEORETICAL BACKGROUND ... 6
Characteristics of Joint-audits ... 6
Audit Quality Determinants ... 8
Individual auditor characteristics and audit quality ... 9
HYPOTHESES DEVELOPMENT ... 10 Firm size ... 10 Gender ... 11 Age ... 12 Tenure ... 12 METHODOLOGY ... 13 Sample selection ... 13 Research design ... 14
Measures of audit quality ... 14
Joint-audit and audit quality ... 16
RESULTS ... 17
Descriptive statistics ... 17
Audit firm size ... 17
3
INTRODUCTION
In 2010, the European Commission suggested that joint-audits could represent a solution to prevent future financial crises by increasing audit quality and by providing a more diverse network of auditing companies additional to the Big4 firms (European Commission, 2010). In the context of this regulatory intent, the French setting is used as a textbook example (European Commission, 2011), since it mandates that listed firms’ financial statements should be subjected to a joint-audit, i.e. being examined by two auditors from different firms, which produce one single audit report. In all other member states of the European Union (EU), an audit done by a single auditor is considered as sufficient verification (Lesage et al., 2017; Holm et al., 2016).1 In this thesis, I address an important omission in the academic literature and draw on a hand-collected dataset from French listed firms to test if in a joint-audit setting pairings of individual auditor characteristics shape the quality of reporting.
The set-up of joint-auditing in the French setting led a lot of research focusing on differentials from other single-audit settings, in terms of costs and benefits. Regarding the former, various studies assess the fees for a joint-audit in comparison to a regular audit assignment (Andre et al., 2016; Deng et al., 2014; Holm et al., 2016). In the Danish setting, the audit fees of a joint-audit were found to be between 10 and 25 percent higher when compared to the ones for a single audit (Lesage et al., 2017). According to Holm et al. (2016), some of the additional fees can be allocated towards coordination costs of an audit pair i.e. crosschecks of the other audit firm’s performance. Since these coordination costs cannot be ignored, this raises questions on whether a joint-audit actually improves the audit quality in spite of the additional fees that it presumes (Andre et al., 2016; Al-hadi et al., 2017; Deng et al., 2014; Lesage et al., 2017). Research on this topic shows that audit fees are significantly higher in France when compared to other countries (Andre et al., 2016) while not necessarily improving
4
audit quality (Andre et al., 2016; Lesage et al., 2017). On the other hand, Zerni et al. (2012) draws on this Swedish sample to show that companies who are voluntarily opting for a joint-audit have a better credit rating, lower abnormal accruals and lower perceived risk i.e. higher audit quality.2 Another important issue that might occur during a joint-audit engagement is the free-riding between a Big4 audit firm and a small audit company. This phenomenon is in turn likely to reduce audit quality and may also increase audit fees (Deng et al., 2014; Holm et al., 2016). Regarding this, extant research shows that in terms of impairments of goodwill, a combination of a Big4 and a non-Big4 firm is more likely to book an impairment which is also larger when impairments occur (Lobo et al., 2017).3 To summarize, existing literature is inconclusive that a joint-audit is favorable over a single audit when looking at audit quality and additional benefits provided by a joint-audit (Al-hadi et al., 2017; Francis, 2006; Lesage et al., 2017; Zerni et al., 2012).
While audit quality and audit fee effects have been widely researched in a joint-audit setting, little is known about the impact of individual traits of auditors taking part into a joint-audit engagement. Specifically, the individual traits of different joint-auditors in a given matched pair may also have influence on the quality of the joint-audit assignment. Nonetheless, most of the extant research is focused on the assessment of individual auditor effects on audit quality only in single-audit setups. For example, in a recent paper, Cameran et al. (2017) researched how audit team composition affects audit quality. Francis (2011) researched the influence of audit teams on audit quality. Albert et al. (2012) shows the relationship between age and risk evasion and Quick and Wiemann (2011) shows a relationship between tenure and audit quality.
2 However, this study also finds that audit fees rise significantly for joint-audits relative to single-audits.
5
In this thesis I expand previous research and evaluate if age, tenure, gender and Big4 affiliation affects audit quality in a joint-audit setting. I build onto an emerging literature stream that analyses the impact of individual auditor characteristics on audit quality of financial reports. To test my contentions, I formulate four sets of expectations. First, I expect that audits performed by at least one Big4 firm to be of higher quality, which is due to a larger firm size and increased reputation risk. I build on Lobo et al. (2017), who find that a combination of one Big4 and one non-Big4 firm provides the highest audit quality. Especially in the joint-audit setup, the collaboration between audit firms is likely to be optimal when mixing Big4 and non-Big4 firms, relative to non-Big4-non-Big4 or non-non-Big4-non-non-Big4 teams. I find that the results of my empirical tests support this contention.
Second, I expect that teams of female individual auditors to perform better quality audits due to a higher risk-aversion and higher levels of independence (Cameran et al., 2017; Hardies et al., 2016). Results show a significant negative relationship between gender and audit quality, which means that teams of male auditors perform higher quality audits. I assign this apparently surprising result to the fact that previous research found female auditors to exhibit a high degree of independence when performing audit assignments. In the joint-audit setting, this would therefore equate with less collaboration between the two audit teams that would likely impair audit quality.
6
auditing procedures. Results show an insignificant relationship between tenure and audit quality.
This paper aims to contribute to the emerging literature stream of joint-audits by narrowing the research gap in the literature which exists today in two areas. First, I expand the literature stream of individual auditor characteristics effects on audit quality (Cameran et al., 2017; Brown et al., 2016; Francis, 2011; Porumb et al., 2017). Second, this thesis adds to our understanding of how joint audit teams can be optimized to produce high quality audits (Lobo et al., 2017; Andre et al., 2016; Zerni et al., 2012). Relative to single-audit setups, joint auditing presumes high levels of collaborative work, which entail particular individual characteristics to play a particularly important role in enhancing audit quality. Overall, my results emphasize the importance of individual characteristics in collaborative setups.
This thesis is build up into various sections. In section 2 a theoretical background is provided which provides relevant theories and knowledge about this research. In section 3 various Hypotheses for this research are formulated. Section 4 provides a methodology which provides information how the data is selected and collected and how the research is going to be conducted. Section 5 provides the results of the various tests to see if the Hypotheses can be accepted or rejected. Various conclusions, possible explanations and possibilities for future research in line with the test results will be drawn in section 7.
THEORETICAL BACKGROUND
Characteristics of Joint-audits
7
be divided between them (Deng et al., 2014). However, the share audit companies take on may be different due to experience or technological advantage of a Big 4 audit firm. At the end both companies check each other’s work and the audit is signed (Deng et al., 2014).
The French audit market is more diverse than other European nations. In 2016 the Big4 companies in France still account for a little less than half of the audit engagements. France has two large mid-tier audit companies named Mazars and Grant Thornton which account for 14 percent of the total market and a wide range of low-tier audit companies or independent auditors which account for 40 percent of the total market. (Audit Analytics, 2016) According to Kermiche et al. (2016), in 2009 Big4 audit engagements accounts for almost 37 percent of the total market. Mid-tier audit firms for 13 percent of the total market and around half of the total market is allocated to independent auditors or low-tier audit firms.
In surrounding countries where a single audit is deemed sufficient, the Big4 market share is larger. In Spain and Italy, the Big4 account for over three quarters of the total market. The Big4 EU average is 60 percent of the total market (Audit Analytics, 2016). It is suggested that a joint-audit diversifies the audit market and provides more space for other, smaller firms (European Commission, 2010).
8
Audit Quality Determinants
Audit quality appears on multiple dimensions and is complex to determine its proxies and their impacts on audit quality (Andre et al., 2016). DeAngelo (1981) defines audit quality as the probability that an auditor would (1) discover an irregularity in the accounting system and (2) report the irregularity. These specific requirements presume that an auditor would be both competent and independent. First, the auditor needs to be competent enough to discover a breach in the accounting principles. Second, the auditor should be able to report the breach which requires independence. In order to measure both competence and independence of an audit report a variety of proxies are required.
One widely used proxy which exists is the amount of accruals. The amount of earnings management is associated with lower audit quality (Lesage et al., 2017; Andre et al., 2016). DeAngelo (1981) state that auditor firm size plays a role in audit quality, arguing that the more auditors within a firm, and thus more or larger clients an audit firm has, leads to higher independence of the audit firm since one client provides for a smaller portion of the total revenue and has less incentives to publish falsified reports to satisfy its clients.Also Dye (1993) argues that larger audit firms have more wealth at risk and therefore have higher incentives to provide high quality reports to prevent litigation by disadvantaged shareholders and creditors of the audited firm shall the audit firm provide low quality of falsified audit reports.
9
Individual auditor characteristics and audit quality
When looking at individual characteristics, the research of Bianchi (2017) with a sample of Italian firms in Italy shows that auditor connectedness is positively associated with audit quality. Companies that are audited by a joint-audit are less likely to restate their financial statements and have lower abnormal accruals. It suggests auditor connectedness leads to knowledge transfer between auditors from different audit firms and therefore leads to higher audit quality.
The paper of Francis (2011) states that the audit teams have influence in the input, process and output of audits and thus influences audit quality. The inputs are mainly based around experience and competence of auditor. The research shows that tenure has a negative influence on audit quality. The argument for this that a long-term relationship has been established between the firm client and the auditor, which makes a critical review of its financial statements more difficult.
Cameran et al. (2017) shows that gender and team diversity also have influence on audit quality by using the sample of two Big4 audit companies in Italy. On average, higher team diversity between partners, managers and audit staff impacts audit quality positively. Also a higher involvement of senior staff and auditors for starting auditors leads to higher quality. The argument given is that higher involvement of senior staff overcomes the lack of knowledge which junior auditors might have. By giving junior employees more attention misstatements occurs more rarely and increases audit quality increases.
10
In recent years the Public Company Accounting Oversight Board (PCAOB) tried to develop an audit quality framework along with indicators for audit quality (PCAOB, 2015). The current framework consists of three dimensions which are professionals, process and results. This can be divided between input (professionals), process and output (results). One of the dimensions to describe individual characteristics is audit professionals. To be specific the experience, skill, training and workload an auditor has. Brown et al. (2016) makes the argument that also gender, firm size and experience influences the audit quality an individual produces. In the research of Brown et al. (2016) a few differences have been reported between gender and firm size when engaging in audit quality improvement processes.
HYPOTHESES DEVELOPMENT
In this section I formulate various hypotheses which synthesize my expectations regarding the impact of the characteristics of audit teams in a joint-audit setting.
Firm size
11
Given this, I expect that a joint-audit assignment which is performed by at least one Big4 firm will result in an audit report of higher quality. I therefore formulate the first Hypothesis:
Hypothesis 1: Audit quality of firms audited by at least one Big4 firm is higher relative
to the audit quality of firms audited by two non-Big4 firms.
Gender
Academic studies assessing the impact of auditor characteristics on audit quality suggest that gender plays an important role in shaping financial reporting. For example, Cameran et al. (2017) shows that a team consisting of at least one female shows higher audit quality. The research of Hardies et al. (2016) on gender impact on audit quality in Belgium shows that females are more likely to issue a going-concern opinion. This effect is larger when the audit report is for an important client. According to Hardies et al. (2016) females are generally more independent and more risk-averse. In addition, female auditors have lower audit errors, which provides decent evidence that females provide higher quality audit reports. According to the research of Cameron et al. (2017) and Hardies et al. (2016) it is expected that females show higher audit quality. In addition, in line with the argumentation of DeAngelo (1981) it is expected that female Big4 auditors perform higher quality audits than non-Big4 auditors. I therefore formulate the following hypotheses:
Hypothesis 2: Audit quality of firms audited by two female auditors is higher than firms
audited by pairs where one or two auditors is male.
Hypothesis 2A: Audit quality of firms audited by two female Big4 auditors is higher
12
Age
Older employees in general have more job experience, which might improve audit quality. According to the research of Albert et al. (2012), where only females were used as research subject, showed that risk aversion increases with age. Specifically, the older an employee is, the higher the chance that this employee evades potential risk vulnerabilities and is more likely to issue an ongoing concern report. In addition, it is expected that with age comes experience. This is due to more audit engagements and more experience and competence in auditing which could lead to higher audit quality. I expect that older auditors perform better quality audits, because of higher risk-aversion (Albert et al., 2012) and more audit engagements. Also in this case it is expected that older Big4 employees generate higher quality audit reports in comparison to older non-big4 employees because of larger reputation risk (DeAngelo, 1981; Dye, 1993).
Hypothesis 3: Audit quality of firms audited by older paired members is higher than
firms audited by younger pairs.
Hypothesis 3A: Audit quality of firms audited by older Big4 audit paired members is
higher than firms audited by older non-Big4 pairs.
Tenure
13
more experience with accepted accounting standards and procedures. Previous literature perceives a positive relationship between audit quality and tenure (Quick et al., 2011; Hohenfels, 2016) I expect that a longer job tenure increases audit quality and that this effect is stronger for Big4 audit firms (Francis, 2009). This leads to hypotheses H4 and H4A.
Hypothesis 4: Audit quality of firms audited by members with a longer tenure is higher
than firms audited by younger pairs.
Hypothesis 4A: Audit quality of firms audited by Big4 pairs with a long tenure is higher
than firms audited by non-Big4 pairs with a long tenure.
METHODOLOGY
Sample selection
I choose to focus on French setting since it is the only one in the European Union (EU) where it is legally required for listed firms to perform a joint-audit (European Commission, 2011). Therefore, listed firms in France represents an ideal choice for conducting my study.
14
auditors or firm was not available or insufficient I delete the firm-year observation. At the end of this process, I obtain a final sample of 1531 auditees, 105 audit firms with 220 offices and 950 individual auditors. In total there are 2,990 observations with available information. Each observation is an audit of a firm-year. Table 1 provides a description of the data.
Table 2 presents the Pearson correlation matrix. The values are generally under 0.7, suggesting that a potential multicollinearity problem in the used variables is reduced.
Research design
Building on previous literature, I use a quantitative approach to construct generally accepted proxies of earnings management. Further, I employ a multiple regression analysis to test the predictions of my hypotheses.
Measures of audit quality
Because audit quality can’t be measured directly, I use various proxies that are developed by previous literature. By describing explanatory power of variables, I can find and understand drivers of audit quality. One method of measurement of audit quality is the amount of abnormal accruals. I measure abnormal accruals for each firm-year. Specifically, I estimate abnormal accruals by utilizing the estimation tools designed by Jones (1991) and further developed by Dechow and Dichew (1995). The Modified Jones Model lowers the amount of error in the measurement of abnormal accruals. I will utilize the long-term accrual model and short-term accrual model.
15
WCACCRx,n = β0,n + β1,n * (ΔREVx,n – ΔARx,n) + εx,n (1)
Where:
WCACCRx,n is calculated as (ΔCAx,n – ΔCASHx,n) - (ΔSTLx,n – ΔLTDx,n – ΔIx,n); REVx,n is the revenue of company x in period n;
ARx,n d is the accounts receivable of company x at time n; CAx,n are the current assets of company x at time n;
CASHx,n is the amount of cash present in company x at time n; STLx,n are the short term liabilities of company x at time n;
LTDx,n provides information about the position of long term debt of the company x at time n; Ix,n stands for the income taxes payable of company x at time n;
All variables are scaled with lagged total assets (TAx,n-1).
From the residuals of the first equation (1) I calculate the short term abnormal working capital accruals as follows:
ABNWCACCRx,n = WCACCRx,n – (b0,n + b1,n * (ΔREVx,n – ΔARx,n)) (2)
All variables are previously defined.
16
TACCRx,n = β0,n + β1,n * (ΔREVx,n – ΔARx,n) + β2,n * GPPEx,n + εx,n (3)
Where:
TACCRx,n is the difference between income before extraordinary items (IBEIx,n) and cash flow from operations (CFOx,n) scaled by lagged total assets. GPPEx,n is gross property, plant and equipment of firm x in period n.
Using the outcome of Equation 3 (TACCx,n), the long term abnormal accruals can be computed as follows:
ABNACCRx,n = TACCRx,n – (b0,n +b1,n * (ΔREVx,n – ΔARx,n) + b2,n * GPPEx,n ) (4)
where:
b0,n; b1,n; and b2,n are regression coefficients which have to be estimated.
The accrual is estimated per industry and year, and I define industry groups by using the industry classification of Barth et al. (1998). I do this since managers have been previously documented to implement their industry-specfic, opportunistic objectives through earnings by using earning accruals as well as cash flow from operations. I therefore use the approach of Barth et al (1998) to control for specific industry accounting practices which are different throughout various industries (Cadman et al., 2013).
Joint-audit and audit quality
17
AQx,n = β0 + β1ONEBIG4 + β2TWOBIGFOUR + β3SIZE + β4ROA + β5SALES_GROWTH + β6LEVERAGE + β7CUR_RATIO + β8LOSS + Year + Industry+ εx,n (5)
where ONEBIG4 is a dummy variable that takes the value of 1 if the firm is audited by one big4 firm and 0 otherwise; TWOBIGFOUR is a dummy variable that has the value of 1 if the firm is audited by at least two big4 firms and 0 otherwise; SIZE is the natural logarithm of total assets; ROA is return of assets and computed by net profit/total assets; SALES_GROWTH is sales growth at time n, computed as (SALESn-SALESn-1)/SALESn-1; LEVERAGE is the ratio between equity on total liabilities; CUR_RATIO is the ratio between current assets and current liabilities; LOSS is a dummy variable which is 1 when there is no profit.
RESULTS
Descriptive statistics
Table 2 shows descriptive statistics of the main variables used in all my estimations.
Audit firm size
18
isolation. Having two Big4 auditors shows insignificant results for BIG4_both (0.044; -0.013). It’s unknown if having two Big4 auditors generate higher quality audits. I accept Hypothesis 1.
Gender
For Hypothesis 2, I expect that females perform higher quality audits. Results in table 5 show strong significant results at Both_Male for MD1 (-0.093 at the 1 percent level) and a weak result for MD2 (-0.073 at the 10 percent level). Additional robustness tests show a more significant relationship for MD2 (-0,073 at the 5 percent level). I reject Hypothesis 2 and claim that a joint-audit of two male joint-auditors perform higher quality joint-audits. This result is surprising since in other audit quality literature females perform higher quality audits (Cameran et al., 2017; Hardies et al., 2016). Previous literature didn’t emphasize on male to male collaboration in a joint-audit setting. I suggest that male to male collaboration is higher and provides better earning management and higher audit quality in comparison to male-female and female-female audit engagements. Since the results are surprising and conflicting with other literature I advise future research into this topic regarding the influence of gender on collaboration in a joint-audit engagement.
Hypothesis 2A in table 6 only shows a positive insignificant result (0,035; 0,005) for MD1 and MD2 respectively. It’s unknown whether having two male Big4 auditors improve audit quality. I reject Hypothesis 2A.
Age
19
reject Hypothesis 3 due to the insignificant result of MD1. It remains unclear if age influences the quality in a joint-audit setting.
In table 8 I show the Hypothesis of Hypothesis 2A. I expect to see a significant negative relationship between age and Big4 affiliation. Results only show insignificant results (0,055 and -0,046) for MD1 and MD2. I reject Hypothesis 3A.
Tenure
For Hypothesis 4 I expect that audit pairs with long tenures perform higher quality audits. I find a slight insignificant positive relationship between audit quality and tenure in table 9 for Both_Experienced (0.031; 0,015) for MD1 and MD2 respectively. I reject Hypothesis 4 due to insignificant results.
Hypothesis 4A in table 10 also shows insignificant results and is rejected (0.050 for MD1 and -0.005 for MD2). It remains unclear if two Big4 auditors with long tenure provide higher quality audits.
ROBUSTNESS
Alternative measures for earnings quality
In line with Porumb et al. (2017) and Andre et al. (2016), I verify the robustness of my results, by running additional tests. I use the method of Dechow et al. (2002) which analyses the relationship between accruals and operational cash flow. The method of Dechow et al. (2002) includes past, present and future cash flows.
20
adjustments to the Jones regression model of total accruals to also include cash flows from the past, present and future.
I calculate my measure of unadjusted total accruals by using the following equation:
WCACCRx,n = b0,n + b1,n * CFOx,n-1+ b2,n * CFOx,n + b3,n * CFOx,n+1 + εx,n
where CFOx,n+x is defined as cash flow from operations for firm x in period n = -1, 0, 1 All results are unchanged when I use this alternative approach.
CONCLUSIONS
Various previous papers evaluated the influences of a joint audit on audit quality or audit fees. Nonetheless, research assessing individual auditor characteristics’ impact on audit quality has been limited and mostly researched in a single auditor setting. In this paper I try to combine both individual auditor characteristics and audit quality in a joint-audit setting.
21
age and tenure with Big4-affiliation were all insignificant. I see no evidence that Big4-auditors have different individual characteristics which could enhance audit quality, which non-Big4 auditors have not.
The findings of this paper have practical implications. The paper provides ways for firms that are subject to a joint-audit to make sure they pair two audit firms with the highest probability of providing a high quality audit, in this case a Big4 and non-Big4 audit firm. Audit firms can select two auditors, each with their individual characteristics, which have the highest chance of providing a high quality audit.
22
23 REFERENCES
Al Hadi, A., Habib, A., Al Yahyaee, K., & Eulaiwi, B. (2017). Joint audit, political connections and cost of debt capital. International Journal of Auditing, 21(3), 249-270.
Albert, S. M., & Duffy, J. (2012). Differences in risk aversion between young and older adults. Neuroscience and Neuroeconomics, 2012(1).
André, P., Broye, G., Pong, C., & Schatt, A. (2016). Are joint audits associated with higher audit fees?. European Accounting Review, 25(2), 245-274.
Audit Analytics. 2016, EU Auditor Market Share: 2016
https://www.auditanalytics.com/blog/eu-auditor-market-share-2016/. Retrieved 1-6-18.
Barth, M. E., Beaver, W. H., & Landsman, W. R. (1998). Relative valuation roles of equity book value and net income as a function of financial health. Journal of Accounting and Economics, 25(1), 1-34.
Bianchi, P. A. (2017). Auditors’ joint engagements and audit quality: evidence from Italian private companies. Contemporary Accounting Research.
Brown, V. L., Gissel, J. L., & Gordon Neely, D. (2016). Audit quality indicators: perceptions of junior-level auditors. Managerial Auditing Journal, 31(8/9), 949-980.
Cadman, B.D., T.O. Rusticus, and J. Sunder. 2013. Stock option grant vesting terms: economic and financial reporting determinants. Review of Accounting Studies 18 (4): 1159–1190.
Cameran, M., Ditillo, A., & Pettinicchio, A. (2017). Audit Team Attributes Matter: How Diversity Affects Audit Quality. European Accounting Review, 1-27.
24
Dechow, P. M., & Dichev, I. D. (2002). The quality of accruals and earnings: The role of accrual estimation errors. The Accounting Review, 77(s-1), 35-59.
Dechow, P. M., Sloan, R. G., & Sweeney, A. P. (1995). Detecting earnings management. Accounting Review, 193-225.
Deng, M., Lu, T., Simunic, D. A., & Ye, M. (2014). Do joint audits improve or impair audit quality?. Journal of Accounting Research, 52(5), 1029-1060.
Dye, R. A. (1993). Auditing standards, legal liability, and auditor wealth. Journal of political Economy, 101(5), 887-914.
European Commission. 2010, Summary Of Responses Green Paper Audit Policy: Lessons From The Crisis http://ec.europa.eu/finance/consultations/2010/green-paper-audit/docs/summary_responses_en.pdf. Retrieved 24-1-18.
European Commission. 2011, Commission Staff Working Paper Impact Assessment
http://ec.europa.eu/internal_market/auditing/docs/reform/impact_assesment_en.pdf. Retrieved 24-1-18.
Francis, J. R. (2011). A framework for understanding and researching audit quality. Auditing: A journal of practice & theory, 30(2), 125-152.
Francis, J. R., and M. D. Yu. 2009. The effect of Big 4 office size on audit quality. The Accounting Review 84 (5): 1521–52.
Francis, J. R., Pinnuck, M. L., & Watanabe, O. (2014). Auditor Style and Financial Statement Comparability. The Accounting Review.
25
Hardies, K., Breesch, D., & Branson, J. (2016). Do (fe) male auditors impair audit quality? Evidence from going-concern opinions. European Accounting Review, 25(1), 7-34.
Hohenfels, D. (2016). Auditor tenure and perceived earnings quality. International Journal of Auditing, 20(3), 224-238.
Holm, C., & Thinggaard, F. (2016). Paying for joint or single audits? The importance of auditor pairings and differences in technology efficiency. International Journal of Auditing, 20(1), 1-16.
Jones, J. J. (1991). Earnings management during import relief investigations. Journal of Accounting Research, 193-228.
Kermiche, L., & Piot, C. (2016). The audit market dynamics in a mandatory joint audit setting– the French Experience. Journal of Accounting, Auditing & Finance, 0148558X16680716.
Lesage, C., Ratzinger-Sakel, N. V., & Kettunen, J. (2017). Consequences of the abandonment of mandatory joint audit: an empirical study of audit costs and audit quality effects. European Accounting Review, 26(2), 311-339.
Lobo, G. J., Paugam, L., Zhang, D., & Casta, J. F. (2017). The effect of joint auditor pair composition on audit quality: Evidence from impairment tests. Contemporary Accounting Research, 34(1), 118-153.
Porumb, V. A., de Jong, A., Huijgen, C., Marra, T. A., & van Dalen, J. (2017). Individual Auditor Style and Audit Quality in a High Reputation Risk Setting.
Public Company Accounting Oversight Board (PCAOB) (2015a), Concept Release on Audit Quality
26 http://
pcaobus.org/Rules/Rulemaking/Docket%20041/Release_2015_005.pdf
Quick, R., & Wiemann, D. (2011). Zum Einfluss der Mandatsdauer des Abschlussprüfers auf die Prüfungsqualität. Zeitschrift für Betriebswirtschaft, 81(9), 915.
27
Table 1
Variable Explanation
MD1 Modified Jones Model
MD2 Long term Modified Jones Model
DD1 Dechow and Dichev model
DD2 Long term Dechow and Dichev model
TAn Total assets at time n
REVn The amount of revenue (or sales) at time n
ARn Accounts receivable at time n
CAn Current assets at time n
CASHn Cash and cash equivalents at time n
STLn Current or short term liabilities at time n
LTDn Current portion of long term debt at time n
In Income taxes payable at time n
GPPEn Gross property plant and equipment at time n
IBEIn Income before extraordinary items at time n
CFOn Cash flow from operations at time n divided by total assets at
time n-1
TACCRn Total accruals at time n, computed as (IBEIn-CFOn)/TAn-1
WCACCRn Working capital accruals at time n
28
ABNACCRx,n Abnormal accruals of company x at time n
ROAn Earnings before interest and taxes at time n/ total assets at time
n-1
SGn Sales growth at time n, computed as (SALESn-SALESn)/SALESn-1
SIZEn Natural logarithm of total assets at time t
DEBTn Total debt to total assets, DEBTn/TAn
29
Table 2 Descriptive statistics
Obs Mean Std. Dev. Min Max
MD_1 3,853 .1494 1039889 1.46e-06 3,952 MD_2 3,637 .0654 .1187 .0000186 4,067 DD_1 3,587 .0678 .1139 .0000157 2,294 DD_2 3,570 .0573 .0942 1.72e-06 2,015 Both_Male 3,853 .4243 .4943 0 1 Both_Old 3,853 .6286 .4832 0 1 Both_Experienced 3,853 .7887 .4082 0 1 BIG4_both 3,853 .0667 .2495 0 1 BIG4_atleastone 3,853 .3122 .4634 0 1 Size 3,794 12,2973 2,3424 4,9628 19,4366 ROA 3,793 -.0018602 .1719 -3347938 12068 Loss 3,853 .2587 .4380 0 1 Leverage 3,791 .2126 .2054 0 3785405
Sales_Growth 3,146 108952.5 1160813 -1.46e+07 4.05e+07
Table 3 Pearson correlation
where MD_1 is the value of the absolute abnormal accruals calculated by the Modified Jones Model; MD_2 is the value of the absolute abnormal accruals calculated by the long term Modified Jones Model; DD_1 is the Dechow and Dichev model; DD_2 is the long term Dechow and Dichev model; Both_male is a dummy variable which takes one when both auditors are male; Both_Old is a dummy variable which has the value of 1 when both auditors are above the average age and 0 otherwise; Both_Experienced is a dummy variable which takes 1 when both auditors have a longer than average and 0 otherwise; BIG4_both is a dummy variable that has the value of 1 if the firm is audited by at least two big4 firms and 0 otherwise; BIG4_atleastone is a dummy variable that takes the value of 1 if the firm is audited by one big4 firm and 0 otherwise; SIZE is the natural logarithm of total assets; ROA is return on assets, computed as net profit/total assets; LOSS is a dummy variable which is 1 when there is no profit; LEVERAGE is the ratio between equity on total liabilities; Sales_Growth is sales growth at time n, computed as (SALESn-SALESn-1)/SALESn-1
MD_1 MD_2 DD_1 DD_2 Both_Male Both_Old Both_Exp BIG4_both BIG4_atleastone Size ROA Loss Leverage Sales_Growth
Table 4
Regression results of Big4 of two Big4 auditors on audit quality (Hypothesis 1)
MD1 MD1 MD2 MD2 BIG4_both 0.044 -0.013 (0.045) (0.049) BIG4_atleast_one -0.135*** -0.136*** (0.032) (0.038) ROA -0.276 -0.346 -0.206 -0.275 (0.293) (0.294) (0.291) (0.294) Loss 0.151** 0.211*** 0.134* 0.195*** (0.068) (0.070) (0.071) (0.074) Leverage -0.201 -0.357* -0.163 -0.318 (0.166) (0.194) (0.207) (0.236) Sales_Growth -0.000 -0.000 -0.000 -0.000** (0.000) (0.000) (0.000) (0.000) Constant 0.807*** 0.012 0.967*** 0.247 (0.147) (0.094) (0.211) (0.169) Observations 2,990 2,990 2,985 2,985 R-squared 0.102 0.088 0.076 0.061
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
32
Table 5
Regression results between gender and audit quality (Hypothesis 2)
MD1 MD2 Both_Male -0.093*** -0.073* (0.035) (0.039) ROA -0.275 -0.203 (0.292) (0.290) Loss 0.147** 0.131* (0.068) (0.071) Leverage -0.208 -0.164 (0.160) (0.201) Sales_Growth -0.000 -0.000 (0.000) (0.000) Constant 0.741*** 0.954*** (0.138) (0.206) Observations 2,990 2,985 R-squared 0.105 0.078
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
Table 6
Regression results of gender, Big4 affiliation and audit quality (Hypothesis 2A)
MD1 MD2 Both_Male -0.097*** -0.074* (0.036) (0.041) BIG4_Both_Male 0.035 0.005 (0.041) (0.049) ROA -0.273 -0.202 (0.292) (0.291) Loss 0.147** 0.131* (0.068) (0.071) Leverage -0.205 -0.164 (0.159) (0.201) Sales_Growth -0.000 -0.000 (0.000) (0.000) Constant 0.752*** 0.956*** (0.142) (0.209) Observations 2,990 2,985 R-squared 0.105 0.078
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
34
Table 7
Regression results impact of age and audit quality (Hypothesis 3)
MD1 MD2 Both_Old 0.045 0.065* (0.031) (0.037) ROA -0.273 -0.199 (0.293) (0.291) Loss 0.149** 0.131* (0.068) (0.071) Leverage -0.207 -0.164 (0.162) (0.201) Sales_Growth -0.000 -0.000 (0.000) (0.000) Constant 0.717*** 0.898*** (0.139) (0.207) Observations 2,990 2,985 R-squared 0.103 0.077
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
35
Table 8
Regression results with age, Big4 affiliation and audit quality (Hypothesis 3A)
MD1 MD2 Both_Old 0.043 0.068* (0.031) (0.037) BIG4_Both_Old 0.055 -0.046 (0.068) (0.062) ROA -0.273 -0.199 (0.293) (0.291) Loss 0.150** 0.130* (0.068) (0.071) Leverage -0.207 -0.165 (0.163) (0.201) Sales_Growth -0.000 -0.000 (0.000) (0.000) Constant 0.727*** 0.891*** (0.140) (0.208) Observations 2,990 2,985 R-squared 0.103 0.078
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
36
Table 9
Regression results of tenure and audit quality (Hypothesis 4)
MD1 MD2 Both_Experienced 0.031 0.015 (0.029) (0.035) ROA -0.276 -0.205 (0.293) (0.291) Loss 0.151** 0.134* (0.068) (0.071) Leverage -0.209 -0.164 (0.165) (0.206) Sales_Growth -0.000 -0.000 (0.000) (0.000) Constant 0.739*** 0.954*** (0.138) (0.213) Observations 2,990 2,985 R-squared 0.102 0.076
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
37
Table 10
Regression results of tenure, Big4 affiliation and audit quality (Hypothesis 4A)
MD1 MD2 Both_Experienced 0.026 0.016 (0.030) (0.036) BIG4_Both_Experienced 0.050 -0.005 (0.049) (0.052) ROA -0.276 -0.205 (0.293) (0.291) Loss 0.151** 0.134* (0.068) (0.071) Leverage -0.205 -0.164 (0.166) (0.206) Sales_Growth -0.000 -0.000 (0.000) (0.000) Constant 0.763*** 0.952*** (0.145) (0.217) Observations 2,990 2,985 R-squared 0.103 0.076
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
