Workforce Environment and Audit Fees (2020)

Journal of Contemporary Accounting and Economics 16 (2020) 100182
Contents lists available at ScienceDirect
Journal of Contemporary
Accounting and Economics
journal homepage: www.elsevier.com/locate/jcae
Workforce environment and audit fees: International evidence

Xuan Sean Sun ⇑, Ahsan Habib, Md. Borhan Uddin Bhuiyan
School of Accountancy, Massey University, Private Bag, 102904 Auckland, New Zealand
a r t i c l e i n f o a b s t r a c t
Article history: Using a dataset from 30 countries over the period from 2002 to 2017, we examine the
Received 7 June 2019 effects of auditing clients’ workforce environment on audit fees as well as the role that
Revised 1 November 2019 national labor market flexibility plays in this relationship. We find evidence that audit fees
Accepted 4 November 2019
are significantly lower for firms with a good workforce environment, suggesting that audi-
Available online 13 January 2020
tors perceive such clients as less risky; as a result, auditors expend less effort and/or charge
a lower risk premium. Furthermore, we find this effect to be stronger for firms in countries
JEL classification:
with a more flexible labor market. Our mediation test results indicate that the relationship
G3
J2
between the audit client workforce environment and audit fees is mediated by media cov-
M5 erage of workforce controversies. Our study contributes to the international audit fee liter-
M14 ature by identifying employee welfare as a distinct audit pricing factor, above and beyond
M42 the effects of overall corporate social responsibility practices.
Ó 2020 Elsevier Ltd. All rights reserved.
Keywords:
Workforce environment
International audit fees
Labor market flexibility
Media coverage
1. Introduction
The auditor is one of the important stakeholders in a modern corporation (Ormazábal, 2018) and understanding how cli-
ent characteristics affect auditors’ behavior is an important and popular area of research in the accounting literature. When
making pricing decisions, auditors evaluate the costs of conducting an audit, the litigation risk inherent therein, and non-
litigation risk brought about by a client’s potential failure (Houston et al., 2005; Pratt and Stice, 1994; Simunic, 1980). Audit
processes and audit efforts are adjusted to reduce any potential costs stemming from client-specific risks. For risks that are
costly to reduce and that cannot be decreased by increasing audit efforts, auditors charge a compensatory premium. There-
fore, any factors related to client-specific risks could influence auditors’ pricing decisions. As employees could be a source of
sustained competitive advantage inasmuch as they facilitate a firm’s success in the market (e.g., Fulmer et al., 2003),
employee- or workforce-related information could be of interest to auditors concerned about clients’ performance and risk.
Auditors have previously voiced their interest in workforce environment. For example, Rebecca Dabbs, a partner of Ernst &
Young, claims that firms can improve workplace productivity by ‘‘. . .better managing environment, health and safety risks.”1
In this paper, we examine whether auditors price information related to a client’s workforce environment. We also investigate
whether country-level labor market characteristics moderate this relationship.
⇑ Corresponding author.
E-mail addresses: [email protected] (X.S. Sun), [email protected] (A. Habib), [email protected] (Md. B.U. Bhuiyan).
1
See https://www.ey.com/en_gl/assurance/how-to-support-better-environment-health-and-safety-outcomes.
https://doi.org/10.1016/j.jcae.2020.100182
1815-5669/Ó 2020 Elsevier Ltd. All rights reserved.
2 X.S. Sun et al. / Journal of Contemporary Accounting and Economics 16 (2020) 100182
Recent studies suggest that a family-friendly workforce environment could not only help firms attract and keep talented
employees (Hom et al., 2017; Turban and Cable, 2003), but also motivate employees to be more cooperative and productive
(Bloom et al., 2011; Fehr and Gächter, 2000). In addition, empirical studies show that relative to peers with a less favorable
environment, firms with a favorable workforce environment have high performance and valuation levels, high levels of inno-
vation performance, few internal control inefficiencies, and a low cost of debt capital, among other advantages (e.g., Chen
et al., 2019, 2016a; Edmans, 2011; Guo et al., 2016; Lee and Kim, 2016). For instance, Guo et al. (2016) find that the prob-
ability of experiencing an employee-related material internal control weakness is relatively low when employees work in a
workforce-friendly environment. In addition, Ji et al. (2018) find that auditors charge higher fees for firms that disclose mate-
rial internal control weaknesses that are non-financial in nature. Therefore, the favorable workforce environment, a key
ingredient of non-financial information, could be interpreted as a signal of both a low level of client internal control risk
and highly credible client financial statements. Such positive signals would decrease audit fees accordingly.
Moreover, a good workforce environment could decrease the incidence of workforce-related controversies. Employees are
less prone to report managerial wrongdoing when they work in a satisfactory environment (e.g., Ben-Nasr and Ghouma,
2018; Rothschild and Miethe, 1999). From an audit risk perspective, less media coverage of workforce controversies is likely
to decrease audit fees, because less public attention on clients will reduce auditors’ litigation risk (e.g., Gong et al., 2018;
Redmayne et al., 2010). Building on these studies, we posit that firms with a favorable workforce environment are relatively
less risky, and auditors consequently charge lower fees, as such clients require both a reduced audit effort and lower audit
risk premiums.
Additionally, the effects of workforce-related information may vary across jurisdictions with different labor market char-
acteristics. Given that labor market flexibility is associated with firing and hiring costs, employee mobility, and worker
behavior (Addison and Teixeira, 2003; Gangl, 2003; Haltiwanger et al., 2014; Ichino and Riphahn, 2005), there is an unre-
solved question as to how labor market flexibility influences the relationship between the workforce environment and audit
fees. In more flexible labor markets, firms with a better workforce environment can easily replace low-productive employees
with more productive ones, as they will have fewer adjustment costs than their peers in less flexible labor markets. At the
same time, in more flexible labor markets, low-productive employees risk losing their job if they shirk their responsibilities,
which will force them to improve their performance. As a result, firms that operate in a flexible labor market and that have a
favorable workforce environment may have stable or even improved employee productivity and, consequently, firm perfor-
mance, thereby reducing audit risk even further. The benefits of investing in the workforce environment are thus likely to be
higher in countries with high labor market flexibility, where firms can easily and quickly adjust their workforce as needed.
Thus, we posit that auditors will further decrease audit fees for firms that have a favorable workforce environment and that
are domiciled in countries with a more flexible labor market.
Using data from publicly listed companies in 30 countries during a sample period from 2002 to 2017, we find that firms
with a favorable workforce environment pay significantly lower audit fees than their counterparts with a relatively poor
workforce environment. The reported coefficient in the baseline model suggests that a one-standard-deviation increase in
the workforce environment index is associated with a 2.57 percent decrease in audit fees. Furthermore, we find that the neg-
ative relationship between the workforce environment and audit fees is strongest in countries with a high level of labor mar-
ket flexibility. Moreover, we find that media coverage of workforce controversies mediates this relationship. Employing a
battery of sensitivity tests, our results remain robust when we use alternative measurements for the workforce environment,
remove clients with multinational operations, control for non-audit services (NAS), and remove countries with an atypical
sample composition. In addition, we use a two-stage-least-squares (2SLS) regression model to alleviate endogeneity con-
cerns stemming from, for example, omitted variables that could influence both the workforce environment and audit fees.
This paper adds to the existing literature in several ways. First, although Huang et al. (2017) document a similar relation-
ship in the U.S.,2 we extend this relationship between the client’s workforce environment and audit behavior (i.e., audit fees) to
an international context, thereby responding to the call for additional, international-level research to exploit the moderating
role of variations in institutional factors on existing findings (e.g., Ball, 2016).
Second, our paper explores issues affecting firms’ operations and performance, rather than employee-related issues asso-
ciated with earnings management (e.g., Chen et al., 2017). Chen et al. (2017) suggest that firms sponsoring defined benefit
pension plans for their employees need more complex accounting estimates, thus providing unique opportunities for man-
agers to manipulate earnings that, in turn, give rise to increased audit fees. Moreover, our study focuses on the supply side
rather than the demand side of audit services. For example, Duellman et al. (2015) find that firms with overconfident man-
agers are less likely to demand a high-quality audit, leading to lower audit fees. We, on the other hand, explore the relation-
ship between the workforce environment and audit fees from the supply-side perspective.
Third, this paper is also distinct from those in the literature that examine the relationship between the voluntary disclo-
sure of corporate social responsibility (CSR) information and audit fees (e.g., Carey et al., 2017; Chen et al., 2016b).
LópezPuertas-Lamy et al. (2017) test the association between firm-level CSR performance and audit fees internationally,
2
Our study also differs from Huang et al. (2017) in two other important ways. First, we examine whether auditors price employer efforts to improve
employees’ working environment, while they rely on employees’ satisfaction with the workforce environment. Second, Huang et al. (2017) do not explore the
possible channel through which employee ratings of their workforce environment might affect audit risk. We address this void by showing that audit clients
with a good workforce environment have fewer workforce-related controversies covered by the media (an outcome that decreases audit fees) relative to their
counterparts with a poor workforce environment.
X.S. Sun et al. / Journal of Contemporary Accounting and Economics 16 (2020) 100182 3
and find evidence that auditors adjust their pricing in response to clients’ CSR performance. Their CSR proxy includes both
social dimensions (e.g., workforce-related variables) and environmental dimensions. However, we incorporate workforce-
related indicators as our main variable of interest for the following reasons. First, in a financial statement audit, an auditor’s
main objective is to be reasonably assured that the financial statements are free of material misstatements owing to fraud or
error. Employees have been found to be the most effective detectors of corporate fraud and misconduct (Dyck et al., 2010);
therefore, workforce-related information is important to auditors. Second, workforce-related information is important not
only to the firms themselves, but also to their business partners. For instance, importers (e.g., Next and H&M) expressed con-
cerns about recent wage disputes in Bangladesh’s garment industry, and requested their auditors to investigate the matter.3
Therefore, an audit client’s workforce environment may have a different impact on audit risk assessment than do other CSR
dimensions.
Our paper proceeds as follows. In Section 2, we present the extant literature on the determinants of audit fees and
employee-related studies, as well as develop our hypotheses. We then describe our research methodology and sample selec-
tion procedure in Section 3. In Section 4, we present descriptive statistics, our main test results, and our mediation test
results. In Section 5, we provide our robustness and endogeneity test results, and we conclude this paper in Section 6.
2. Literature review and hypotheses development
Academic studies suggest that the costs of conducting an audit (i.e., audit fees) consist of a ‘‘resource cost factor” and an
‘‘expected loss factor” (e.g., Pratt and Stice, 1994; Simunic, 1980). The resource cost factor is a quantitative measurement of
how many audit efforts the auditor performed, while the expected loss factor is the present value of future losses for which
the auditor will be liable, which is related to client-specific risks. Furthermore, auditors will expend their audit efforts to
reduce future expected losses to the point at which the overall cost of conducting an audit is expected to be the lowest. Audi-
tors charge a litigation risk premium to compensate for the remaining expected loss factors (Simunic and Stein, 1996). More-
over, Houston et al. (2005) introduce a non-litigation risk premium into the audit fee model. In sum, audit efforts, litigation
risk, and non-litigation risk premiums collectively comprise the audit fees that auditors charge.
According to Hay et al. (2006), most of the determinants of audit fees can be categorized into client attributes, auditor
attributes, and engagement attributes. We consider audit clients’ workforce-related information as a client attribute that will
affect an auditor’s pricing behavior. The role of employees in modern corporations has been debated for several years.
Instead of considering employees as a common input factor in the production process, modern management theory suggests
that employees are a resource crucial to a firm’s success. According to the resource-based view, firms’ resources are classified
into three categories: physical capital resources, human capital resources, and organizational capital resources (Barney,
1991). Human capital resources consist of training, experience, judgment, intelligence, relationships, and the insights of in-
dividual managers and workers in a firm. Human capital resources meet the criteria for sustained competitive advantage
because of their VRIN characteristics (i.e., valuable, rare, inimitable, and non-substitutable), which distinguish human capital
resources from other types and help firms to pursue competitive success (Pfeffer and Villeneuve, 1994; Wright et al., 1994).
Therefore, corporations must necessarily improve and maintain their employees’ effectiveness and efficiency to remain
competitive. In addition to giving normal monetary incentives, employers can motivate employees by also providing non-
monetary, welfare-related benefits. These non-monetary benefits include opportunities to be involved in the firm’s
decision-making process, a flexible working schedule, a safer workplace environment, and more training and career devel-
opment opportunities, among others. Employees make more effort and perform better in their tasks when they work in a
good workforce environment, which leads to higher productivity and performance (Akerlof, 1982; Bloom et al., 2011;
Dalal et al., 2012; Levine, 1992). Meanwhile, employees are also more cooperative and less likely to become involved in sab-
otaging activities when they are treated well (Fehr and Gächter, 2000). Many other studies, both in single country and inter-
national settings, present evidence that firms with satisfied employees are more likely to outperform counterparts with
dissatisfied employees (e.g., Chen et al., 2016a; Edmans, 2011, 2012; Faleye and Trahan, 2011; Fauver et al., 2018; Filbeck
and Preece, 2003; Gupta and Krishnamurti, 2018; Huselid, 1995; Lee and Kim, 2016). Firms with a high employee-
friendly rating tend to maintain low leverage ratios (Bae et al., 2011), and debtholders require lower returns (Chen et al.,
2019). Note that these benefits may be limited to treating employees well within an appropriate range, and that excess
employee welfare treatment could have detrimental effects on shareholder value (Ben-Nasr and Ghouma, 2018).
On the other hand, Hom and Kinicki (2001) suggest that in response to an unsatisfactory workforce environment, employ-
ees may choose to either leave a firm or to perform in a detrimental way. Therefore, poor employee treatment policies sig-
nificantly increase the likelihood of an employee-related material internal control weakness and the probability of a
misstatement caused by an unintentional error (Guo et al., 2016). Firms with a poor workforce environment are more likely
to have employee disputes (e.g., strikes and litigation) that lead to significant financial and reputational losses (e.g., Karpoff
and Lott, 1999). Thus, auditors will increase their professional skepticism toward client-specific risks when a client has a
poor workforce environment, a proposition that has been empirically confirmed in U.S. contexts by Huang et al. (2017).
Given these studies, the client’s workforce environment may affect audit fees in multiple ways. First, a better workforce
environment may lead to a lower likelihood of a financial statement misstatement, thereby reducing audit efforts and hence
3
See https://qz.com/1540275/5000-garment-workers-in-bangladesh-were-fired-after-protesting-low-wages/amp.
audit fees. Employees who work in a favorable environment are more likely to properly perform internal tasks and hence
significantly reduce the propensity for having material internal control weaknesses. Second, the litigation risk premiums that
auditors charge may also be lower when clients have a favorable workforce environment, because auditors are less likely to
be sued for material misstatements. Relatedly, such firms tend to have fewer workforce-related disputes and controversies
reported in the media, leading to less public attention on both the firms and their auditors, as well as decreased litigation
risks for the auditor. Overall, firms with a good workforce environment are highly likely to be seen to have relatively few
audit risks; therefore, auditors will expend less effort on such clients and charge them lower risk premiums. We therefore
state our first hypothesis as follows:
H1: A good (poor) audit client’s workforce environment decreases (increases) audit fees, ceteris paribus.
H1 focuses on the firm-specific workforce environment determinants of audit fees. However, prior studies document that
institutional factors (e.g., the broader legal environment) also play a vital role in determining audit fees (e.g., Choi et al.,
2008; Kuo and Lee, 2016; Taylor and Simon, 1999). Our second hypothesis considers country-level labor market flexibility,
referring to the speed with which labor adapts to fluctuations and changes in market conditions, as a possible moderator
between the client’s workforce environment and audit fees. Labor market flexibility has been found to be negatively asso-
ciated with the restrictiveness of employment protection regulations (e.g., Addison and Teixeira, 2003; Gangl, 2003;
Haltiwanger et al., 2014). Employment protection regulations influence firms’ employee-related costs including per worker
employment costs and employment adjustment costs (Addison and Teixeira, 2003). The former reflects the costs of hiring and
providing benefits to employees, and the latter reflects costs that accompany gross changes, especially those that occur when
employers dismiss employees. Both costs will be higher in countries/markets with more protective employment regulations
and, consequently, lower labor market flexibility. In such countries, firms are less likely to hire and fire employees, which
decreases both employees’ labor mobility and their external opportunities (Gangl, 2003). In countries with a more flexible
labor market, high external job availability allows employees to switch firms more easily.
Moreover, when more outside options are available, firms need to pay higher compensation to retain key skilled employ-
ees. According to the theoretical discussion from Shapiro and Stiglitz (1984), the total demand for labor declines when all
firms raise their wages. Thus, employees who shirk their responsibilities are under the threat of job loss, which encourages
them to make more firm-specific investments. Ichino and Riphahn (2005) use data from a large Italian bank and document
that the number of days of absence per week, on average, increases significantly after employees are protected by the
employment regulation. Bjuggren (2018) uses a natural experiment method in Sweden to explore the causal relationship
between labor market flexibility and labor productivity. He finds that after a reform of Swedish last-in-first-out (LIFO) labor
rules,4 labor productivity increased by 2 percent to 3 percent in small treated firms compared to large firms that were not
affected by the reform. The author suggests that the increased threat of job loss may induce employees to exert greater effort
in their jobs. Edmans et al. (2017) and Gupta and Krishnamurti (2018) provide further evidence on the benefits of undertaking
employee-friendly practices in countries with high labor market flexibility.
Building on this discussion, we conjecture that in flexible labor markets, firms with a favorable workforce environment
are more likely to retain productive employees and to lay off low-productivity employees, thus creating for incumbent
employees a job termination threat that forces them to work diligently. Therefore, firms with a good workforce environment
should maintain or increase their employees’ performance easily, because of the low hiring and firing costs in flexible labor
markets. Therefore, the benefits brought about by a good workforce environment will be enhanced in countries with a more
flexible labor market. As a result, auditors’ concerns about such firms’ specific risk will further diminish with increased labor
market flexibility. Accordingly, we state our second hypothesis as follows:
H2: The impact of the workforce environment on audit fees will be reinforced in countries with high labor market flexibility.
3. Research methodology
3.1. Model specification
To test H1, we develop the following Ordinary Least Square (OLS) regression model:
LANF i;t ¼ b0 þ b1 WEI i;t þ b2 SIZEi;t þ b3 LEV i;t þ b4 INVREC i;t þ b5 ROAi;t
þb6 LOSSi;t þ b7 NBSi;t þ b8 NGSi;t þ b9 SPECIALi;t þ b10 CROSSi;t þ b11 MTBi;t
þb12 CURRENT i;t þ b13 INTSi;t þ b14 TURN i;t þ b15 ISSUEi;t þ b16 CSRi;t
ð1Þ
þb17 BIGNi;t þ b18 AOi;t þ b19 BUSY i;t þ b20 BSIZEi;t þ b21 BINDi;t þ b22 ACM i;t
þb23 ACMINDi;t þ b24 ACMEXP i;t þ b25 CEODUALi;t þ b26 LAW i;t þ b27 LNGDP i;t
þb28 FDIi;t þ b29 DISCLi;t þ b30 CORRUP i;t þ Fixed Effects þ ei;t ;
where the dependent variable is the natural logarithm of audit and audit-related fees (LNAF) (see Section 3.2 for a detailed
construction). Our variable of primary interest is the Workforce Environment Index (WEI), which provides a comprehensive
4
Before 2001, regulations mandated that if the firm wished to lay off staff, the employee who last joined the firm should be the first fired. After the 2001
reform, firms with <11 employees were granted the option to choose which of the three shortest-tenure employees should be let go.
rating of a company’s workforce environment (see Section 3.3 for a detailed discussion). In H1, we argue that firms with a
good workforce environment pay significantly lower audit fees. Therefore, we predict a negative b1.
We include several control variables that are likely to determine audit fees. Firm size (SIZE), measured as the natural log-
arithm of total assets, is found to be an extremely critical explanatory variable for audit fees (Hay et al., 2006). In addition, we
use five variables to control for firm complexity, including INVREC (the sum of inventories and receivables divided by total
assets), NBS (the natural logarithm of the number of business segments), NGS (the natural logarithm of the number of geo-
graphic segments), TURN (net sales divided by total assets), and INTS (coded as 1 for firms that have at least 10 percent inter-
national sales to total sales, and zero otherwise). Since operational or geographical diversification and internationalization
often signal greater complexity in a firm’s operations, such firms require more audit efforts and procedures, resulting in
higher audit fees (e.g., Choi et al., 2008; Jaggi and Low, 2011; Kim et al., 2012). Thus, we expect positive associations between
audit fees and firm size and the complexity measures.
We control for several client-specific risk factors because auditors will either make greater auditing efforts or charge fee
premiums to high-risk clients, leading to increased audit fees (e.g., Pratt and Stice, 1994). Consistent with Choi et al. (2008),
proxies for the client-specific risks are LEV, ROA, LOSS, SPECIAL, MTB, and CURRENT. We measure LEV as the sum of short- and
long-term debt divided by total assets, ROA as net income divided by total assets, and LOSS as a dummy variable that is coded
as 1 for firms reporting negative income before extraordinary items for the current year, and zero otherwise. Firms that
report special items (SPECIAL) are coded as 1, and zero otherwise. The market to book ratio (MTB) is measured as the ratio
of firm market capitalization to common shareholder equity, and the liquidity ratio (CURRENT) is measured as total current
assets divided by total current liabilities. We predict positive coefficients on LEV, LOSS, and SPECIAL, and negative coefficients
on ROA, MTB, and CURRENT. We use cross-listing on the U.S. markets (CROSS) as an additional variable for client-specific lit-
igation risk, because auditors face increased legal liability when client firms are cross-listed in countries with stronger legal
regimes than those in their home country (Choi et al., 2009). Similar to Kuo and Lee (2016), we include firms’ financing activ-
ities (ISSUE) as a control variable that is coded as 1 when a firm obtains either equity or debt capital in the current year, and
zero otherwise; we include this particular control variable since both the demand for audit quality and audit risk are higher
when firms are involved in such activities. We also include CSR as an additional control variable.
We also include a set of auditor and engagement attributes likely to affect audit fees, including BIGN auditor (a dummy
variable coded as 1 for firms audited by a Big N audit firm, 0 otherwise), audit opinion (AO; a dummy variable coded as 1 for
firms receiving qualified opinions, 0 otherwise), and busy season (BUSY; a dummy variable coded as 1 for firms for which the
fiscal year-end comes during an auditor busy season, 0 otherwise).5 All these audit-related variables are expected to have a
positive relationship with audit fees. Carcello et al. (2002), Hay et al. (2008), and Zaman et al. (2011) find that firms with a good
corporate governance structure are more likely to demand high-quality audits, whereas other studies suggest the opposite (e.g.,
Griffin et al., 2008). As proxies for a client’s corporate governance structure, we include board size (BSIZE; the number of board
members), board independence (BIND; the proportion of independent board members), CEO duality (CEODUAL; a dummy vari-
able coded as 1 for firms in which the CEO and chair of the board are the same person, 0 otherwise), the presence of an audit
committee (ACM; a dummy variable coded as 1 for firms with an audit committee, 0 otherwise), audit committee independence
(ACMIND; the proportion of independent audit committee members), and audit committee expertise (ACMEXP; a dummy vari-
able coded as 1 for firms that have an audit committee with at least three members and at least one financial expert, 0
otherwise).
Similar to Choi et al. (2009) and Kuo and Lee (2016), we include five country-level control variables that may affect cross-
country variations in audit fees. LAW is a dummy variable for legal origin that is coded as 1 for common law countries and 0 for
code law countries, which represents a country’s legal origin and the level of investor protection; FDI is the level of foreign
direct investment as a proportion of gross domestic product (GDP); LNGDP is the natural logarithm of GDP per capita; DISCL
measures a country’s required disclosure level; and CORRUP is the perceived corruption index. We expect all the country-level
variables to be positively related to audit fees. We further include year- and industry-fixed effects in our Eq. (1).
In H2, we hypothesize that the negative relationship between the workforce environment and audit fees is stronger in
countries with high labor market flexibility. We develop the following Eq. (2) to test H2:
LNAF i;t ¼ b0 þ b1 WEIi;t þ b2 LMF i;t þ b3 LMF i;t WEI i;t þ Control v ariables þ Fixed Effects þ ei;t ; ð2Þ
where LMF refers to the country-level labor market flexibility (for a detailed construction of LMF, see Section 3.4). In Eq. (2),
our main variable of interest is the interaction term between LMF and WEI (i.e., LMFi, * WEIi,t), and we expect a negative asso-
ciation between LNAF and this interaction term (b 3 < 0.) Such a negative association would indicate that auditors tend to
further decrease fees for client firms with a favorable workforce environment, as the country-level labor market flexibility
increases.
3.2. Measurement of audit fees
We use the natural logarithm of the sum of audit fees and audit-related fees (LNAF) as our proxy for audit pricing, as is
consistent with U.S. audit fees research. The most common measure of audit fees in international audit fees research
5
Prior studies use a single month (e.g., December or January) to proxy the auditor busy season effect for all countries. In this study, however, for each country
in our initial sample, we designate the month when the largest number of firms have their fiscal year-end as the auditor busy month.
(e.g., Bronson et al., 2017; Choi et al., 2008) is ‘‘fees paid to auditor” and includes (1) audit and audit-related fees and (2)
other non-audit-services fees. We did not include the fees paid for non-audit services because regulations permitting or pro-
hibiting the provision of such services vary widely across jurisdictions. Consequently, we use audit and audit-related fees,
rather than the total fees paid to the auditor, as our proxy for audit pricing.6
3.3. Measure of workforce environment index
We retrieve firm-level workforce environment data from the Thomson Reuters ESG database, which covers information
related to ESG (Environmental, Social, and Governance) and significant ESG controversies.7 The Thomson Reuters ESG data-
base provides over 400 ESG measures and both enhances and replaces the ASSET4 database used in prior studies (e.g., Gupta and
Krishnamurti, 2018; Thomson Reuters, 2018). Following Gupta and Krishnamurti (2018), we identify a list of 20 indicators that
are available consistently across the sample period to construct the Workforce Environment Index (WEI).8 These indicators com-
bine both qualitative (19) and quantitative (1) attributes. We provide the details of each measure and the scoring approach we
use in Appendix A. We add all of a firm’s scores to construct a WEI that ranges from 0 to 20. A high value for WEI indicates that a
firm has a favorable workforce environment.
3.4. Measurement of labor market flexibility
We use two measures of labor market flexibility. The first is the Employment Protection Legislation (EPL) Index for OECD
countries and some emerging countries. This index measures ‘‘the procedures and costs involved in dismissing individuals or
groups of workers and the procedures involved in hiring workers on fixed-term or temporary work agency contracts.”9 It
consists of three category scores: individual dismissals of regular workers (EPR), additional costs of collective dismissals
(EPC), and the regulation of temporary contracts (EPT). Consistent with prior studies (Banker et al., 2013; Pagano and Volpin,
2005), we measure EPL as the equally weighted EPR, EPC, and EPT. Furthermore, we multiply EPL by 1 for ease of interpreta-
tion. Following Edmans et al. (2017), we use the labor market regulations index (EFW) from the Economic Freedom of the World
database as our second measure of labor market flexibility to represent the ‘‘de facto strictness of labor regulation” (Feldmann,
2009, p. 77). There are six components of EFW: hiring regulations and minimum wage (5Bi), hiring and firing regulations (5Bii),
centralized collective bargaining (5Biii), working-hours regulations (5Biv), mandated cost of worker dismissal (5Bv), and con-
scription (5Bvi). We use the average of these six components to construct the EFW index. High values for both EPL and EFW
indicate greater labor market flexibility.10
3.5. Sample
We extract audit and audit-related fees data, all of which required firm-specific financial information from the Thomson
Reuters Fundamentals (via Thomson Reuters Eikon) and Worldscope databases. We retrieve workforce-related information
from the Thomson Reuters ESG database, as we discussed in Section 3.3. Auditor classification information is gathered from
Thomson Reuters Eikon, which provides time-series auditor information. We source firm-level governance data from the
Thomson Reuters ESG database, and country-level variables mainly from World Bank, the International Monetary Fund
(IMF), and prior literature (e.g., La Porta et al., 2008). The country-level required disclosure index is developed by the Center
for International Financial Analysis and Research (CIFAR). The perceived corruption index, developed by Transparency Inter-
national, ranges from 1 to 10. For ease of interpretation, we replace the original index with 10 minus the original value,
which means that the higher the index, the higher the perceived corruption level. In Appendix B, we include detailed variable
definitions and sources. The sample period is from 2002 to 2017. We choose 2002 as the beginning year, because workforce-
related information became available beginning in that year.
We begin with a sample of 32,666 firm-year observations with non-missing audit fees and WEI data from 43 countries for
the 2002–2017 sample period. Consistent with previous literature, we eliminate firm-years from the financial (4787) and
utility industries (1323) based on the 2-digit Global Industry Classification Standards (GICS), as the audit fee determinants
for such industries are distinct. After further applying the data requirements for computing firm-specific financial variables,
6
Audit-related fees include fees paid for external services that are reasonably associated with performing the audit or reviewing financial statements. We
find similar results when we use only audit fees as the dependent variable.
7
The Thomson Reuters ESG database provides ESG data for 7000+ companies globally, which gives us a platform for comparing ESG performance globally.
The ESG database collects ESG information from multiple sources, including annual reports, NGO websites, CSR reports, company websites, stock exchange
fillings, and news sources, among others.
8
Gupta and Krishnamurti (2018) identify 35 workforce-related performance indicators to construct their own employee treatment index. Because the
Thomson Reuters ESG database is an updated version of the ASSET4 database and it classifies some indicators that are better related to the workforce, we
identify 20 measures among those indicators to capture the overall workforce environment. However, we acknowledge that our WEI based on the selected
indicators may not capture the entire workforce environment. Therefore, in Section 5.1, we confirm the validity of WEI using alternative independent variables
gathered from the ESG and the ASSET4 databases.
9
This definition is retrieved from the OECD website: http://www.oecd.org/employment/emp/oecdindicatorsofemploymentprotection.htm.
10
The current form of the EPL data is available annually until 2013, and the EFW data is available annually until 2016. To maximize our sample size in Eq. (2),
we replace the missing EPL and EFW using their latest values.
Table 1
Industry distribution.
Industry group Name N % of N

1010 Energy 1791 7.93
1510 Materials 3011 13.34
2010 Capital Goods 3361 14.89
2020 Commercial & Professional Services 828 3.67
2030 Transportation 1110 4.92
2510 Automobiles & Components 656 2.91
2520 Consumer Durables & Apparel 877 3.89
2530 Consumer Services 828 3.67
2540 Media 736 3.26
2550 Retailing 1234 5.47
3010 Food & Staples Retailing 375 1.66
3020 Food, Beverage & Tobacco 982 4.35
3030 Household & Personal Products 277 1.23
3510 Health Care Equipment & Services 906 4.01
3520 Pharmaceuticals, Biotechnology & Life Sciences 1001 4.43
4510 Software & Services 1206 5.34
4520 Technology Hardware & Equipment 1044 4.62
4530 Semiconductors & Semiconductor Equipment 548 2.43
5010 Telecommunication Services 568 2.52
6010 Real Estate 1234 5.46
Total 22,573 100.00
we delete 3876 observations. Similar to Jaggi and Low (2011), we drop some countries (Brazil, Chile, Greece, Hungary,
Indonesia, Luxembourg, Mexico, Philippines, Portugal, and Thailand) with <30 firm-year observations. Our baseline regres-
sion, therefore, includes 22,573 firm-year observations from 30 countries. We then delete 2454 firm-year observations with
missing corporate governance variables. Finally, we exclude 315 observations with missing country-level data, resulting in
19,804 firm-year observations that we use for the most comprehensive regression specification.11 To mitigate the impact of
outliers, we winsorize all the continuous variables at the 1st and 99th percentiles of their respective distributions. In Table 1, we
present our industry-based sample distributions. Our sample is well distributed, as no single industry covers more than 15 per-
cent of the total sample. A total of 14.89 percent of the sample comes from Capital Goods (GICS 2010), followed by 13.34 percent
from Materials (GICS 1510) and 7.93 percent from Energy (GICS 1010).
4. Results
4.1. Descriptive statistics
In Table 2, we report the detailed descriptive statistics of the variables used in this study. As shown in Panel A, the mean
of LNAF is 7.44, corresponding to 3.91 million $US, which is similar to the result found by LópezPuertas-Lamy et al. (2017).12
The average workforce environment index is 10.37, exhibiting a large standard deviation. The mean SIZE of the sample firms is
15.23 (i.e., 12,362 million $US), while about 14 percent of firms report negative incomes for the current year, and the majority of
them (INTS statistic of 0.70) have substantial international operations. The ratio of receivables and inventory over total assets is
23 percent, while current assets are, on average, two times larger than current liabilities. The mean leverage ratio is 55 percent,
while 8 percent of sample firms are cross-listed in the U.S. markets. With respect to the firm-level governance variables, we note
that the mean size of the board of directors is 10. Most of the firms (93 percent) have an audit committee, which shows that
most countries follow a best practice code that suggests the formation of an audit committee. The percentages of independent
directors on the board of directors and the audit committee are 57 percent and 80 percent, respectively. Moreover, about 75
percent of audit committees have at least three members and at least one financial expert. Big N audit firms audit most of
the sample firms (94 percent), and only a few firms (0.2 percent) received qualified audit opinions. In addition, about 74 percent
of firms have their fiscal year-end during the auditor’s busy season.
In Table 2, Panel B, we provide country distribution and descriptive statistics for our main variables (i.e., LNAF and WEI)
and the country-level variables. About 31 percent of the sample comes from the United States (6960 observations), while
around 24 percent is from Japan (2932 observations) and the United Kingdom (2495 observations). Russia (30 observations)
has the lowest representation followed by Israel (39 observations) and Ireland (71 observations). The mean LNAF and WEI
reveal significant variation across countries. For example, firms are more likely to have a better workforce environment
when they are from certain European countries (e.g., France, Germany, Spain, Sweden). Regarding the country-level variables
11
Our sample is comparable to Fauver et al. (2018) who use 21,103 observations to investigate the impact of employee-friendly culture on firm value.
12
The average amount of audit fees is calculated using unlogged audit fees, which, for our comprehensive sample observations, is 3.98 million $US.
Table 2
Descriptive statistics.
Panel A: Firm-level variables

Variable N Mean SD P25 Median P75
LNAF 22,573 7.44 1.32 6.53 7.44 8.34
WEI 22,573 10.37 4.00 7.00 11.00 14.00
SIZE 22,573 15.23 1.52 14.30 15.22 16.18
LEV 22,573 0.55 0.22 0.41 0.55 0.68
INVREC 22,573 0.23 0.16 0.10 0.22 0.34
ROA 22,573 0.05 0.10 0.02 0.05 0.09
LOSS 22,573 0.14 0.35 0.00 0.00 0.00
NBS 22,573 1.02 0.71 0.00 1.10 1.61
NGS 22,573 1.19 0.70 0.69 1.39 1.79
SPECIAL 22,573 0.88 0.32 1.00 1.00 1.00
CROSS 22,573 0.08 0.28 0.00 0.00 0.00
MTB 22,573 3.01 3.47 1.18 2.05 3.66
CURRENT 22,573 1.92 1.47 1.09 1.52 2.24
INTS 22,573 0.70 0.46 0.00 1.00 1.00
TURN 22,573 0.86 0.59 0.45 0.76 1.13
ISSUE 22,573 0.27 0.44 0.00 0.00 1.00
CSR 22,573 51.62 17.40 38.17 51.59 65.05
BIGN 20,119 0.94 0.23 1.00 1.00 1.00
AO 20,119 0.00 0.05 0.00 0.00 0.00
BUSY 20,119 0.74 0.44 0.00 1.00 1.00
BSIZE 20,119 9.73 3.10 8.00 9.00 11.00
BIND 20,119 0.57 0.28 0.38 0.62 0.80
ACM 20,119 0.93 0.26 1.00 1.00 1.00
ACMIND 20,119 0.80 0.34 0.67 1.00 1.00
ACMEXP 20,119 0.75 0.43 0.00 1.00 1.00
CEODUAL 20,119 0.40 0.49 0.00 0.00 1.00
LAW 19,804 0.71 0.46 0.00 1.00 1.00
LNGDP 19,804 10.64 0.55 10.61 10.76 10.89
FDI 19,804 0.04 0.08 0.02 0.02 0.03
DISCL 19,804 76.14 5.42 74.00 76.00 80.00
CORRUP 19,804 2.49 1.03 2.40 1.90 2.70
EPL 18,366 1.60 0.52 1.83 1.50 1.13
EFW 19,804 8.06 1.19 7.70 8.33 9.10
Panel B: Country distribution (N=22,573)
Country N % of N LNAF WEI LAW LNGDP FDI DISCL CORRUP EPL EFW
Australia 1849 8.19 6.307 9.981 1 10.93 0.0046 80 1.82 1.79 7.64
Austria 100 0.44 6.871 12.06 0 10.78 0.0188 62 2.48 2.31 5.94
Belgium 116 0.51 7.136 12.01 0 10.70 0.0631 68 2.53 3.14 7.27
Canada 1336 5.92 6.999 9.240 1 10.75 0.0390 75 1.61 1.38 8.31
China 353 1.56 6.839 9.181 0 8.883 0.0101 – 6.11 2.42 5.55
Denmark 198 0.88 7.399 11.80 0 10.97 0.0264 75 0.85 2.14 7.40
Finland 218 0.97 7.338 12.42 0 10.77 0.0269 83 0.97 1.79 5.33
France 642 2.84 8.703 13.88 0 10.62 0.0263 78 3.01 3.13 5.70
Germany 679 3.01 7.992 13.78 0 10.69 0.0315 67 2.03 2.46 6.18
Hong Kong 709 3.14 7.119 9.395 1 10.54 0.3220 73 2.17 – 9.33
India 259 1.15 5.858 11.81 1 7.365 0.0044 61 6.28 1.85 6.83
Ireland 71 0.31 7.397 9.775 1 10.93 0.2310 81 2.56 1.81 7.79
Israel 39 0.17 7.758 8.641 1 10.44 0.0230 74 3.92 1.60 5.29
Italy 168 0.74 8.039 13.14 0 10.45 0.0123 66 5.56 2.88 6.68
Japan 2932 12.99 7.569 10.64 0 10.61 0.0257 71 2.49 1.83 8.21
Malaysia 245 1.09 6.283 12.13 1 9.205 0.0413 79 5.17 – 7.95
Netherlands 257 1.14 8.386 12.62 0 10.82 0.3440 74 1.43 2.29 7.01
New Zealand 135 0.60 5.768 8.948 1 10.57 0.0003 80 0.91 0.81 8.72
Norway 135 0.60 7.586 11.57 0 11.37 0.0416 75 1.41 2.61 4.77
Poland 99 0.44 5.899 9.364 0 9.491 0.0102 – 4.14 2.29 7.47
Russia 30 0.13 9.009 9.533 0 9.346 0.0282 – 7.41 1.90 5.84
Singapore 325 1.44 6.567 9.606 1 10.79 0.1210 79 1.16 – 7.65
South Africa 503 2.23 7.009 13.07 1 8.769 0.0142 79 5.65 1.51 6.21
South Korea 454 2.01 6.154 10.80 0 10.15 0.0210 68 4.58 2.12 4.70
Spain 253 1.12 7.544 13.42 0 10.30 0.0335 72 3.89 2.83 5.54
Sweden 395 1.75 7.598 12.25 0 10.90 0.0436 83 1.06 1.99 6.39
Switzerland 303 1.34 7.840 11.93 0 11.29 0.0762 80 1.32 2.12 8.01
Taiwan 315 1.40 5.508 10.97 0 9.990 0.0002 58 3.91 – 5.90
United Kingdom 2495 11.05 7.035 11.53 1 10.65 0.0280 85 2.04 1.45 8.26
United States 6960 30.83 8.131 8.778 1 10.86 0.0209 76 2.61 1.13 9.12
Note: Variables are defined in Appendix B.

shown in Table 2, over 71 percent of firms come from common law countries; the average disclosure requirement is over 76,
and the mean perceived corruption index is about 2.49, suggesting that the countries covered in our sample provide high
levels of protection and financial information to investors in general. Also, labor markets are more flexible in Hong Kong
(mean EFW is 9.33), the United States (mean EFW is 9.12), and New Zealand (mean EFW is 8.72), and they are less flexible
in South Korea (mean EFW is 4.70) and Norway (mean EFW is 4.77).
4.2. Correlation analysis
Table 3 presents the Pearson correlations among the variables included in our baseline regression. The test variable WEI is
significantly and negatively correlated with LNAF (correlation coefficient of 0.20, p < 0.01), indicating that a good workforce
environment decreases audit fees in the absence of relevant controls. This univariate result supports our prediction. Consis-
tent with the existing auditing literature, we find that audit fees (LNAF) are positively and significantly correlated with SIZE
(correlation coefficient of 0.73), LEV (coefficient of 0.32), complexity (e.g., coefficients of 0.41 and 0.35 for NBS and INTS,
respectively), and BIGN (coefficient of 0.22, untabulated). The positive correlations between firm-level governance variables
(i.e., BSIZE, BIND, ACM, ACMIND, and ACMEXP) and LNAF indicate that well-governed firms are likely to demand more audit
services (untabulated). Regarding country-level variables, audit fees are higher in countries with good economic conditions
(coefficient of 0.21 for LNGDP, untabulated).
4.3. Workforce environment and audit fees: Baseline regression results for H1
In Table 4, we report the OLS regression results for H1. Throughout this paper, we compute all reported t-statistics in
parentheses using heteroscedasticity-robust standard errors clustered by firm. In Column (1), we report the association
between audit fees (LNAF) and the workforce environment index (WEI), controlling for financial statement-based control
variables. In Column (2), we include firm-level corporate governance variables, and Column (3) introduces country-level con-
trol variables. In addition, we control for firms’ CSR performance in Column (4) to test whether workforce-related perfor-
mance impacts audit fees after controlling for overall CSR practice performance. The coefficients on WEI are consistently
negative and statistically significant for all the specifications mentioned above. These results support H1, suggesting that
auditors may expend less effort, and charge lower risk premiums, in response to a better audit client’s workforce environ-
ment. In terms of economic significance, the reported coefficient in Column (4) suggests a decrease of US$ 102,292 in audit
fees for a one-standard-deviation increase in WEI.13 Given that the unlogged mean value of audit fees is US$ 3.98 million, a
one-standard-deviation increase in WEI would decrease the mean audit fees by 2.57 percent. Therefore, the effects of the work-
force environment on audit fees are both statistically and economically significant.
With respect to the sign and significance of the control variables, we find that large and complex firms (e.g., SIZE, NBS,
NGS, TURN, INTS) and high audit risk firms (e.g., high LEV) are more likely to pay relatively high audit fees. The coefficients
on BIGN and AO are positive and significant, indicating that there is a Big N premium and a risk premium for high audit risk
firms. Consistently, we find that BSIZE, BIND, ACM, and ACMEXP are positive and significant, suggesting that good corporate
governance structures are more likely to demand a high-quality audit. Also, the coefficients for LAW, LNGDP, DISCL, and COR-
RUP are positive and significant at the 1 percent level, as expected. Taken together, our findings suggest that the client’s
workforce environment plays an important role in the audit pricing process.
We conduct a diagnostic test to mitigate possible concerns about multicollinearity. Our untabulated result shows that all
the variance inflation factors (VIFs) are well below 10 (Marquaridt, 1970) (the highest value is 4.14 for ACMIND). Therefore,
multicollinearity should not be a concern in this study.
4.4. Workforce environment and audit fees: moderation results for H2
In this section, we test whether audit fees are related to the interactive effects of a country’s labor market flexibility and
WEI. H2 predicts that the effect of the workforce environment on audit fees will be stronger in countries with a more flexible
labor market. We report our results in Table 5, Columns (1) (EPL*WEI) and (2) (EFW ⁄ WEI). In both columns, our variable of
interest (i.e., WEI) is still negatively and significantly associated with audit fees, supporting our main results (coefficients of
0.010 (p < 0.05) and 0.011 (p < 0.01) on WEI, respectively). Furthermore, the coefficients of EPL ⁄ WEI and EFW ⁄ WEI are
negative and statistically significant at the 1 percent level in Columns (1) and (2), respectively.14 Thus, our results suggest
13
Following Kuo and Lee (2016), we first multiply the coefficients (excluding year and industry) from Table 4, Column (4) with the mean values of the
corresponding variables as reported in Table 2. The sum of these numbers equals the logged audit fees (LNAF) of 6.98, which implies:
e6:98 1000 ¼ US1; 074; 918. Then, we recalculate the sum by replacing the mean value of WEI, 10.37, with 14.37, where 14.37 equals a one-standard-
deviation increase in the mean value of WEI, all else being the same. Thus, the revised LNAF is 6.88, while the unlogged audit fees are US$ 972,626. The
differences are $972,626-$1,074,918=US$ 102,292.
14
In our main test, we replace missing EPL and EFW with their latest values. If we use only the available EPL and EFW data, then our sample sizes decrease to
9994 for EPL and 18,406 for EFW, respectively. Our results remain the same if we use smaller samples. Moreover, to mitigate the multicollinearity problem
resulting from the introduction of interactive terms and to ease interpretation, we apply a mean-centering approach to both the WEI and EPL (EFW) variables
before constructing the interaction terms.
10
X.S. Sun et al. / Journal of Contemporary Accounting and Economics 16 (2020) 100182
Table 3
Correlations.
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16)
LNAF (1) –
WEI (2) 0.20 –
SIZE (3) 0.73 0.40 –
LEV (4) 0.32 0.16 0.27 –
INVREC (5) 0.06 0.13 0.05 0.14 –
ROA (6) 0.02 0.06 0.08 0.10 0.13 –
LOSS (7) 0.07 0.08 0.15 0.04 0.13 0.64 –
NBS (8) 0.41 0.20 0.39 0.12 0.12 0.07 0.11 –
NGS (9) 0.25 0.14 0.16 0.03 0.12 0.02 0.03 0.16 –
SPECIAL (10) 0.06 0.15 0.09 0.08 0.01 0.08 0.05 0.05 0.06 –
CROSS (11) 0.05 0.05 0.04 0.09 0.09 0.12 0.10 0.07 0.02 0.03 –
MTB (12) 0.00 0.05 0.13 0.09 0.03 0.23 0.08 0.06 0.04 0.11 0.07 –
CURRENT (13) 0.21 0.20 0.29 0.52 0.00 0.06 0.10 0.11 0.01 0.07 0.08 0.01 –
INTS (14) 0.35 0.21 0.22 0.02 0.16 0.05 0.04 0.20 0.41 0.09 0.00 0.00 0.02 –
TURN (15) 0.09 0.04 0.09 0.19 0.57 0.21 0.15 0.05 0.03 0.07 0.12 0.15 0.16 0.00 –
ISSUE (16) 0.02 0.05 0.01 0.03 0.05 0.09 0.06 0.05 0.00 0.00 0.05 0.00 0.01 0.03 0.09 –
Note: Sample size: 22,573. Boldface indicates significance at the 1% level. Variables are defined in Appendix B. This table reports the correlations between the variables used in the baseline regression. Other
untabulated correlations are generally as expected, except for the insignificant correlation between AO and FDI.
Table 4
Baseline regression: Workforce environment and audit fees.
(1) (2) (3) (4)

LNAF LNAF LNAF LNAF
WEI 0.009** 0.010*** 0.012*** 0.023***
(2.38) (3.14) (3.66) (5.17)
SIZE 0.587*** 0.548*** 0.567*** 0.558***
(56.24) (51.77) (52.79) (51.32)
LEV 0.813*** 0.589*** 0.495*** 0.501***
(11.06) (9.97) (8.68) (8.73)
INVREC 0.899*** 0.349*** 0.030 0.015
(8.12) (3.08) (0.24) (0.12)
ROA 0.807*** 0.925*** 0.795*** 0.787***
(7.88) (9.82) (8.75) (8.69)
LOSS 0.085*** 0.029 0.025 0.021
(3.25) (1.19) (1.05) (0.91)
NBS 0.196*** 0.183*** 0.177*** 0.173***
(8.97) (8.92) (9.16) (8.88)
NGS 0.148*** 0.108*** 0.095*** 0.094***
(8.54) (6.77) (6.03) (5.98)
SPECIAL 0.092*** 0.004 0.041 0.044
(3.12) (0.14) (1.52) (1.64)
CROSS 0.106** 0.126*** 0.065 0.056
(2.37) (2.96) (1.49) (1.29)
MTB 0.016*** 0.005* 0.007** 0.006**
(5.04) (1.94) (2.46) (2.24)
CURRENT 0.058*** 0.034*** 0.015* 0.014*
(6.22) (4.28) (1.95) (1.88)
INTS 0.391*** 0.370*** 0.386*** 0.385***
(13.89) (13.83) (14.11) (14.05)
TURN 0.430*** 0.312*** 0.250*** 0.247***
(13.74) (10.49) (8.07) (8.00)
ISSUE 0.010 0.012 0.009 0.010
(0.73) (0.91) (0.71) (0.77)
CSR – – – 0.003***
(3.50)
BIGN – 0.411*** 0.229*** 0.228***
(7.69) (4.40) (4.38)
AO – 0.338*** 0.261** 0.264**
(2.59) (2.18) (2.18)
BUSY – 0.033 0.078*** 0.080***
(1.27) (2.98) (3.08)
BSIZE – 0.029*** 0.034*** 0.034***
(6.72) (7.91) (7.92)
BIND – 0.821*** 0.638*** 0.603***
(13.67) (10.32) (9.73)
ACM – 0.137** 0.223*** 0.222***
(2.29) (4.01) (4.02)
ACMIND – 0.078 0.356*** 0.363***
(1.22) (5.75) (5.86)
ACMEXP – 0.180*** 0.104*** 0.101***
(6.39) (3.96) (3.88)
CEODUAL – 0.165*** 0.131*** 0.136***
(7.44) (5.93) (6.12)
LAW – – 0.171*** 0.187***
(3.82) (4.18)
LNGDP – – 0.469*** 0.470***
(11.33) (11.37)
FDI – – 0.335* 0.290
(1.87) (1.62)
DISCL – – 0.023*** 0.023***
(6.71) (6.76)
CORRUP – – 0.138*** 0.140***
(6.18) (6.29)
Industry Yes Yes Yes Yes
Year Yes Yes Yes Yes
_cons 3.189*** 3.864*** 10.779*** 10.744***
(19.71) (22.86) (20.08) (20.01)
N 22,573 20,119 19,804 19,804
Adj.R-Square 0.64 0.71 0.74 0.74
Note: Robust t-statistics are in brackets. Variables are defined in Appendix B.

***p < 0.01, **p < 0.05, *p < 0.10.
Table 5
Interactive effects of the workforce environment and labor market flexibility on audit fees (Moderation effects).
(1) (2)
LNAF LNAF
WEI 0.010** 0.011***
(2.39) (2.70)
EPL 0.165*** –
(3.71)
EPLWEI 0.041*** –
(7.29)
EFW – 0.208***
(12.41)
EFWWEI – 0.016***
(6.91)
Other control variables SIZE; LEV; INVREC; ROA; LOSS; NBS; NGS; SPECIAL,
CROSS, MTB, CURRENT, INTS, TURN, ISSUE, CSR, BIGN,
AO, BUSY, BSIZE, BIND, ACM, ACMIND, ACMEXP,
CEODUAL, LAW, LNGDP, FDI, DISCL, CORRUP
Industry Yes Yes
Year Yes Yes
_cons 8.576*** 11.065***
(15.65) (21.10)
N 18,366 19,804
Adj.R-Square 0.77 0.75

***p < 0.01, **p < 0.05, *p < 0.10.
that the impact of WEI on audit fees is stronger in countries with a more flexible labor market, where employees have more
outside opportunities and higher productivity. The findings of our empirical tests support H2.
4.5. Mediation test results
The mediation effect refers to ‘‘the presence of an intervening variable or mechanism that transmits the effect of an ante-
cedent variable on an outcome” (Aguinis et al., 2017, p. 666). Our results thus far do not identify the specific channel(s)
through which the workforce environment reduces audit fees. We propose the financial reporting quality (FRQ) and media
coverage of workforce controversies (CONTRO) as two such possible channels. We choose FRQ because Guo et al. (2016) pro-
vide evidence that firms with an employee-friendly environment are less likely to have employee-related material internal
control weaknesses and financial restatements. Poor FRQ has been found to increase audit fees (e.g., Cho et al., 2017; Ji et al.,
2018). We use a number of FRQ measures to conduct our mediation test.15 We choose CONTRO as another mediating variable
because employees are less prone to report managerial wrongdoing when they work in a satisfactory environment. Therefore, a
negative association is expected between the client firm’s workforce environment and the frequencies of CONTRO. CONTRO may
increase audit fees because either the negative media coverage provides more information related to client firms’ risks to audi-
tors (i.e., the information role), or media coverage places more public attention on clients, which will increase an auditor’s lit-
igation risks (i.e., the disciplining role) (e.g., Gong et al., 2018; Redmayne et al., 2010).
Following Baron and Kenny (1986), we use the following four steps to establish these mediation channels (Eqs. (3A)–
(3C)). First, we show that variations in the independent variable (i.e., WEI, in our study) are correlated with the dependent
variable (i.e., audit fees, LNAF; Eq. (3A); a1), to confirm the possibility that a mediation effect is present. Second, we show
how variations in the independent variable (i.e., WEI) account for variations in the mediator (i.e., M; Eq. (3B); b1). Third,
we show that the mediator (i.e., M) has a significant effect on the dependent variable (i.e., LNAF; Eq. (3C); c2). Finally, we
show that the significant relationship between WEI and LNAF (Eq. (3A)) either becomes insignificant after controlling for
M (full mediation) or that the significance level shrinks after doing so (partial mediation). To conduct our mediation test,
we develop the following set of equations:
LNAF i;t ¼ a0 þ a1 WEIi;t þ Control v ariables þ Fixed Effects þ ei;t ; ð3AÞ
Mi;t ¼ b0 þ b1 WEIi;t þ Control v ariables þ Fixed Effects þ ei;t ; ð3BÞ
LNAF i;t ¼ c0 þ c1 WEIi;t þ c2 Mi;t þ Control v ariables þ Fixed Effects þ ei;t : ð3CÞ
15
Owing to data unavailability at the international level for clients’ material internal control weaknesses and financial restatements, we were unable to
perform a mediation test using these two variables. Instead, in this section, we employ several FRQ measures suggested by the prior literature, such as
discretionary accruals (Dechow et al., 1995; Kothari et al., 2005), real earnings management (Roychowdhury, 2006), and earnings smoothness (Dechow et al.,
2010). We find consistent results for all these FRQ measures. For brevity’s sake, we report only results for the discretionary accruals test.
The mediators (i.e., M) are |DAC| and CONTRO. |DAC| is measured as the absolute value of discretionary accruals generated
from the modified Jones model (Dechow et al., 1995), while CONTRO is measured as the natural logarithm of the total num-
ber of controversies published in the media related to workforce plus one.16 The total effect of WEI on LNAF (i.e., a1 from Eq.
(3A)) can be decomposed into a direct effect and an indirect effect (i.e., through the mediator). The direct effect is c1 from Eq.
(3C), whereas the indirect effect is b1 ⁄ c2 for the proposed mediator. The core of the mediation effect rests in testing the null
hypothesis H0: b1 ⁄ c2 = 0. For the estimation we provide, we first use the OLS regressions to perform Baron and Kenny’s (1986)
causal step regression. However, we acknowledge that Baron and Kenny’s (1986) causal step regression has come under criti-
cism in recent years (Hayes, 2009; Zhao et al., 2010). Zhao et al. (2010) suggest a superior test approach (i.e., the ‘‘bootstrap”
approach proposed by Preacher and Hayes, 2004) for the mediation test. Therefore, we report the results of both Baron and
Kenny’s (1986) causal step regression and the bootstrap approach.17
We report the direct and indirect effects of the workforce environment on audit fees for the pooled sample in Table 6,
Panels A and B. In Columns (1) and (4) of Panel A, we document negative and significant coefficients on WEI when we
run the regressions without mediators (i.e., |DAC| and CONTRO; Eq. (3A)), which is in line with our H1 that a good audit cli-
ent’s workforce environment decreases audit fees. We do not find a significant relationship between WEI and |DAC| in Col-
umn (2), indicating that a client’s workforce environment has no direct impact on the overall FRQ for our chosen FRQ
measures.18 Therefore, we conclude that our chosen FRQ proxies do not mediate the relationship between the workforce envi-
ronment and audit fees. Our result in Column (3) suggests that our main result remains unchanged, even after controlling for
the FRQ in the model. On the other hand, we find a significantly negative association between WEI and CONTRO in Column (5)
(coefficient 0.003, p < 0.05), suggesting that the frequency of media coverage of workforce controversies is lower for firms with
a better workforce environment. As we show in Column (6), there is a partial mediation effect between WEI and LNAF through
CONTRO. The results from the bootstrap approach support Baron and Kenny (1986)’s causal step regression results by showing
significant indirect effects for CONTRO but not for |DAC|. However, the direct effects account for the bulk of the total effects. In
sum, our evidence suggests that media coverage of workforce controversies, a hitherto unexplored channel in the international
audit fee literature, affects the relationship between the workforce environment and audit fees.
5. Additional robustness tests
5.1. Alternative measure of workforce environment
So far in this paper, we use the workforce-related measures from the ESG database to construct our workforce environ-
ment index (WEI). Although this approach is prevalent in the prior literature (e.g., Fauver et al., 2018; Gupta and
Krishnamurti, 2018), we now examine whether we obtain consistent results when we employ alternative workforce envi-
ronment (WEI_ALT) proxies. The first alternative variable is the Workforce Score (WS) calculated and provided by the ESG
database, which measures ‘‘. . .a company’s effectiveness towards job satisfaction, a healthy and safe workplace, maintaining
diversity and equal opportunities, and development opportunities for its workforce” (Thomson Reuters, 2018, p. 15). WS is
benchmarked to the Thomson Reuters Business Classification (TRBC) industry groups. Thus, the higher the WS, the better a
firm’s workforce environment is, when compared to its industry peers. Furthermore, as the WS is a substitute for four sub-
scores that are not benchmarked to industry groups in the ASSET4 database,19 we use the equal weighted workforce score
(WS_A4) as our second alternative variable by taking the average of those subscores. We re-estimate Eqs. (1) and (2) using
WS and WS_A4 as the independent variables. Our results, reported in Table 7, Panel A, show negative relationships between
the workforce environment (both WS and WS_A4) and audit fees, and thus support H1. Moreover, such relationships are stronger
in countries with a more flexible labor market, which support H2. Overall, as we show in Table 7, Panel A, our findings are robust
to alternative proxies for the workforce environment. These consistent results also validate the construction of our WEI.
5.2. Exclusion of multinational firms
In our main analyses, we add INTS (coded as 1 for firms that have at least 10 percent international sales to total sales, and
zero otherwise) to control for the impact of internationalization on audit pricing. However, the internationalization of audit
16
The Thomson Reuters ESG database provides the following controversy indicators for the workforce: (1) the number of controversies published in the media
linked to workforce diversity and opportunity; (2) the number of controversies published in the media linked to workforce health and safety; (3) the number of
controversies published in the media linked to the company’s relations with employees or relating to wage or wage disputes; and (4) whether an important
executive management team member announced a voluntary departure (other than retirement) or had been ousted. We add up the three quantitative indicators,
(1) to (3), to derive the total number of controversies published in the media related to workforce. In addition, the original WEI includes an indicator, ‘‘Is the
company under the spotlight of the media because of a controversy linked to the company’s employees, contractors or suppliers due to wage, layoff disputes or
working conditions?,” which is related to our media coverage proxy. Thus, we exclude this indicator and re-construct our WEI for this analysis.
17
The bootstrap approach is a non-parametric method based on resampling with replacement, which is done multiple times (e.g., 5000 times in our study).
For a more detailed discussion of the use of the bootstrap approach in mediation tests, see Hayes (2018).
18
We also find insignificant relationships between WEI and other FRQ measures from Eq. (3B): performance-matched discretionary accruals (coefficient
0.001, p = 0.32), real earnings management (coefficient 0.014, p = 0.551), and earnings smoothness (coefficient 0.007, p = 0.377). We caution readers that
our mediation results are based on the chosen FRQ measures and that we could not rule out the possibility that other, unselected FRQ measures might mediate
the relationship between the workforce environment and audit fees.
19
Those scores are employment quality (SOEQ), health and safety (SOHS), training and development (SOTD), and diversity and opportunities (SODO).
Table 6
Regression results of mediation tests (Mediation effects).
Panel A: The Baron and Kenny’s (1986) causal step regression results
Dependent Variable (1) (2) (3) (4) (5) (6)
LNAF |DAC| LNAF LNAF CONTRO LNAF
(3A) (3B) (3C) (3A) (3B) (3C)
WEI 0.023*** 0.001 0.025*** 0.022*** 0.003** 0.022***
(5.17) (0.85) (5.42) (5.08) (2.56) (5.04)
|DAC| – – 0.018* – – –
(1.76)
CONTRO – – – – – 0.068**
(2.24)
Other controls Yes Yes Yes Yes Yes Yes
Industry and year Yes Yes Yes Yes Yes Yes
_cons 10.744*** 1.252*** 10.989*** 10.778*** 0.962*** 10.712***
(20.01) (9.93) (18.66) (20.04) (8.52) (19.90)
N 19,804 17,574 17,574 19,804 19,804 19,804
Adj.R-Square 0.74 0.10 0.74 0.74 0.13 0.74
Panel B: Bootstrap approach results
Direct effect 0.0247*** 0.0223***
Indirect effect 0.0000 0.0002***
Total effect 0.0247*** 0.0225***
Indirect/Total 0.0009 0.0089
Indirect/Direct 0.0009 0.0090
Total/Direct 1.0009 1.0090
Note: Our primary independent variable is WEI, and our dependent variable is LNAF. We use as mediators financial reporting quality proxied by discre-
tionary accruals (i.e., |DAC|) and media coverage of workforce controversies (CONTRO). Column (1) shows the effect of the independent variable (WEI) on the
dependent variable (LNAF) without the mediator (|DAC|). Column (2) presents the effect of WEI on |DAC|, while Column (3) shows the results with both the
independent variable (WEI) and the mediator (|DAC|). Columns (4) to (6) report the results of the mediation test using CONTRO as the mediator. Robust t-
statistics are in brackets. Variables are defined in Appendix B.
***p < 0.01, **p < 0.05, *p < 0.10.
clients may also influence their workforce environment. More concretely, multinational audit clients must adopt the best
standards from all the countries in which they operate. As a result, their workforce environment affected by all of these coun-
tries’ employment regulations, which may introduce noise in our main findings. To mitigate this issue, we partition our sam-
ple into multinational (INTS = 1; 13,942 observations) and domestic clients (INTS = 0; 5864 observations). The results
(untabulated) show that the mean value of WEI in our multinational clients subsample (i.e., 10.92) is significantly higher
than that in our domestic clients subsample (i.e., 9.12), supporting the prediction that to comply with multiple countries’
employment regulations, multinational clients offer a relatively better workforce environment. We re-run our Eqs. (1)
and (2) using the domestic clients subsample only, and the results (untabulated) show significantly negative coefficients
on WEI in Eq. (1) and on EPL ⁄ WEI (EFW ⁄ WEI) in Eq. (2), which is consistent with our main findings.
5.3. Control for non-audit services fees
We include an additional variable (LN_NAS, the natural log of the sum of tax and other services fees) in our equations to con-
trol for the possible association between audit fees and NAS fees, based on the notion that the simultaneous provision of audit
and NAS may reduce audit costs owing to either economies of scope or knowledge spillover (e.g., Chung and Kallapur, 2003;
O’Keefe et al., 1994). Our untabulated results show that our main variable of interest (i.e., WEI) is still significantly and nega-
tively associated with LNAF across all specifications, which supports our main findings. In contrast to our expectations, the coef-
ficient on LN_NAS is significantly positive. A possible explanation suggested by Simunic (1984) is that knowledge spillover helps
auditors reduce the unit costs of audit services and, in turn, the reduced price encourages clients to buy additional audit
services.
5.4. Additional analyses
We also perform two additional analyses. First, as we show in Table 2, Panel B, about 33 percent of the observations in our
sample are from the U.S., leading to a concern that U.S. firms might be primarily driving our findings. To mitigate this con-
cern, we re-run our models by excluding all U.S. observations. Our results in Table 7, Panel B suggest that our results do not
suffer from sample concentration. Second, there is a concern that auditors may decide how much they will charge at the
beginning of the year, which may lead to incorrect inferences when we use the workforce environment from the current year
to predict audit fees for those current years. We mitigate this concern by repeating Eqs. (1) and (2) using lagged WEI and
Table 7
Robustness and the endogeneity test.
Panel A: Alternative independent variables

WS WS_A4
(1) (2) (3) (4) (5) (6)
LNAF LNAF LNAF LNAF LNAF LNAF
WEI_ALT 0.297*** 0.138*** 0.175*** 0.328*** 0.206** 0.131*
(5.75) (2.79) (3.56) (4.68) (2.85) (1.98)
EPL – 0.158*** – – 0.202*** –
(3.63) (3.94)
EPLWEI_ALT – 0.529*** – – 0.631*** –
(6.97) (6.63)
EFW – – 0.201*** – – 0.201***
(12.23) (10.61)
EFWWEI_ALT – – 0.195*** – – 0.242***
(6.24) (6.31)
Control variables Yes Yes Yes Yes Yes Yes
Industry Yes Yes Yes Yes Yes Yes
Year Yes Yes Yes Yes Yes Yes
_cons 11.122*** 8.589*** 10.874*** 9.661*** 7.967*** 10.100***
(20.56) (15.66) (20.60) (15.37) (11.94) (16.21)
N 19,804 18,368 19,804 15,839 14,945 15,839
Adj.R-Square 0.74 0.77 0.75 0.75 0.78 0.77
Panel B: Non-US results and results with lagged independent variables

Non-US observations Lagged independent variables
(1) (2) (3) (4) (5) (6)
LNAF LNAF LNAF LNAF LNAF LNAF
WEI 0.010* 0.012* 0.011* 0.023*** 0.013*** 0.012***
(1.75) (1.96) (1.90) (4.79) (2.78) (2.68)
EPL – 0.302*** – – 0.137*** –
(5.65) (2.95)
EPLWEI – 0.023*** – – 0.038*** –
(2.89) (6.68)
EFW – – 0.128*** – – 0.192***
(6.85) (11.18)
EFWWEI – – 0.012*** – – 0.015***
(4.03) (5.86)
Control variables Yes Yes Yes Yes Yes Yes
Industry Yes Yes Yes Yes Yes Yes
Year Yes Yes Yes Yes Yes Yes
_cons 7.209*** 5.749*** 8.223*** 9.830*** 7.930*** 10.263***
(11.25) (8.83) (12.60) (17.24) (13.47) (18.26)
N 13,292 11,854 13,292 16,576 15,347 16,576
Adj.R-Square 0.72 0.76 0.73 0.73 0.76 0.74
Panel C: 2SLS results for the endogeneity test
(1) (2)
WEI LNAF
WEI – 0.044**
(2.22)
COUN_WEI 0.272*** –
(9.73)
IND_WEI 0.374*** –
(16.60)
Control variables Yes Yes
Industry Yes Yes
Year Yes Yes
_cons 9.310*** 10.840***
(7.18) (20.26)
N 19,786 19,786
(Centered) R-Square 0.726 0.739
Kleibergen-Paap rk LM statistics 280.989
(under-identification test) (p=0.0000)
Kleibergen-Paap rk Wald F statistic (weak identification test) 259.312
Hansen J statistic (over-identification test) 2.073
Note: Robust t-statistics are in brackets. Variables are defined in Appendix B. All the independent variables used in Panel B, Columns (4) to (6) are lagged
forms.
***p < 0.01, **p < 0.05, *p < 0.10.
lagged independent variables. The coefficients on lagged WEI and on EPL ⁄ WEI (EFW ⁄ WEI) are also significantly negative, as
we report in Panel B of Table 7. These results further support our main findings.
5.5. Endogeneity test
We further address endogeneity concerns about omitted variables by implementing a two-stage-least-squares (2SLS)
regression. We report our results in Table 7, Panel C. Similar to normal CSR practice (Ioannou and Serafeim, 2012), we argue
that a firm’s workforce environment is determined by both country and industry characteristics. Ioannou and Serafeim
(2012, 2017) suggest that in a given country, a firm’s CSR practice might vary over time systematically as a result of regu-
latory changes. In addition, a firm’s incentives and abilities to engage in CSR are influenced by its competitors in the same
industry (Hawn and Kang, 2018). If a firm operates in an industry in which its peers have a stronger commitment to
employee well-being and that firm cannot provide a workforce environment comparable with those of its competitors, then
it risks losing employees. Thus, firms are more likely to provide a favorable workforce environment if they are headquartered
in countries, as well as operate in industries, that have a stronger commitment to employee well-being. Following prior stud-
ies (e.g., Cheng et al., 2014; Gupta and Krishnamurti, 2018), we select the country-year mean WEI (COUN_WEI) and the
country-industry mean WEI (IND_WEI) as our instruments. When calculating both COUN_WEI and IND_WEI, we exclude
the focal firm’s workforce environment performance to eliminate the firm’s influence on both instrumental variables. As a
result, our instruments represent the average workforce environment of the focal firm’s competitors within the same indus-
try and across years, in a given country.20 We predict positive effects for COUN_WEI and IND_WEI on WEI. However, we have no
a priori reason to believe that COUN_WEI and IND_WEI to have a direct impact on audit fees through channels other than the
firm-level workforce environment.
In the first stage, we regress WEI on COUN_WEI, IND_WEI and other control variables that we used in Eq. (1). Our results in
Column (1) of Table 7, Panel B show that both the coefficients on COUN_WEI and IND_WEI are positive and significant at the 1
percent level, supporting our argument that a firm’s workforce environment is determined by both country and industry
characteristics. In the second stage, we replace the original WEI with the predicted WEI generated from the first stage.
Our results in Column (2) of Table 7, Panel B show that the coefficients on WEI remain negative and significant at the 5 per-
cent level, further supporting our main finding that a favorable workforce environment leads to lower audit fees. Similar to
Cheng et al. (2014), we perform three tests to check the validity of our instruments in Table 7, Panel B. First, the result of the
under-identification test, the Kleibergen-Paap rk LM statistic, illustrates that our model is identified (p = 0.0000). Second, we
report the Kleibergen-Paap rk Wald F statistic (Kleibergen and Paap, 2006) for our weak identification test. The F-statistic is
very high in our sample, suggesting that our instruments are relevant and strong. Third, we report the Hansen’s J statistic
(Hansen, 1982) to test the over-identification concern. The p-value of the Hansen’s J statistic is high for this test, suggesting
that we cannot reject the null hypothesis that the instruments are exogenous. Following Larcker and Rusticus (2010), we also
regress the residuals of the second stage on the exogenous variables (i.e., COUN_WEI, IND_WEI, and control variables) to test
the over-identification concern. Our untabulated results support the Hansen’s J statistic. Overall, our post-estimation tests
confirm both the relevance and the exclusion restrictions of our instruments.
5.6. Alternative argument on workforce environment
Our results suggest that auditors charge lower audit fees for client firms that provide a favorable workforce environment.
However, this negative relationship may hold up only to a certain point, as this relationship may reverse if a client firm
improves its workforce environment beyond that point. From the agency perspective, opportunistic managers may provide
generous welfare to employees to increase employee satisfaction, which will reduce the likelihood that employees report
managers’ wrongdoing. From a traditional cost efficiency perspective (e.g., Taylor, 1914), employers should keep the employ-
ees from being too satisfied by paying them no more than their reservation wages because over-satisfaction might lead to
complacency and shirking responsibilities. Based on both agency and efficiency perspectives, auditors may assess higher
risks and, in turn, increase auditing efforts and audit fees, when client firms’ employee welfare is too generous. Therefore,
a U-shaped relationship may exist between the workforce environment and audit fees.
We introduce a quadratic term, WEI2, into our Eq. (1) and report our results in Table 8. The coefficient on WEI remains
negative and significant. The positive and significant coefficient of WEI2 confirms that the optimal level (i.e., WEI = 18.98)
for a workforce environment is very close to the maximum value of our workforce environment index (20).21 Our results sug-
gest that auditors charge higher fees only to those clients with an extremely favorable workforce environment, to compensate
for the extra risks related to agency or employee efficiency problems. However, the untabulated results find that this U-shaped
relationship exists only in firms with poor corporate governance mechanisms, indicating that good corporate governance mech-
anisms can mitigate auditors’ concerns about the extreme workforce environment.
20
We acknowledge the limitation inherent in selecting industry-average as the instrument (Larcker and Rusticus, 2010). Larcker and Rusticus (2010) suggest
that accounting researchers often use regulatory changes as a quasi- experiment to address endogeneity concerns. However, owing to the international nature
of our study, we could not identify a universal regulatory shock.
21 coefficient on WEI
We derive the optimum level of 18.98 using the formula: 2coefficient on WEI2
¼ ð0:0491243Þ 2
20:0012939 ¼ 18:98. Both coefficients on WEI and WEI are from Table 8,
Column (3).
Table 8
Additional test results: Non-linear relationship between the workforce environment and audit fees.
(1) (2) (3)

LNAF LNAF LNAF
WEI 0.034*** 0.037*** 0.049***
(2.60) (3.00) (3.82)
WEI2 0.001* 0.001** 0.001**
(1.83) (2.05) (2.14)
Other control variables SIZE; LEV; INVREC; ROA; SIZE; LEV; INVREC; ROA; SIZE; LEV; INVREC; ROA;
LOSS; NBS; NGS; LOSS; NBS; NGS; LOSS; NBS; NGS;
SPECIAL, CROSS, MTB, SPECIAL, CROSS, MTB, SPECIAL, CROSS, MTB,
CURRENT, INTS, TURN, CURRENT, INTS, TURN, CURRENT, INTS, TURN,
ISSUE, BIGN, AO, BUSY, ISSUE, BIGN, AO, BUSY, ISSUE, CSR, BIGN, AO,
BSIZE, BIND, ACM, BSIZE, BIND, ACM, BUSY, BSIZE, BIND, ACM,
ACMIND, ACMEXP, ACMIND, ACMEXP, ACMIND, ACMEXP,
CEODUAL CEODUAL, LAW, LNGDP, CEODUAL, LAW, LNGDP,
FDI, DISCL, CORRUP FDI, DISCL, CORRUP
Industry Yes Yes Yes
Year Yes Yes Yes
_cons 3.754*** 10.654*** 10.613***
(21.14) (19.73) (19.65)
N 20,119 19,804 19,804
Adj.R-Square 0.71 0.74 0.74

***p < 0.01, **p < 0.05, *p < 0.10.
6. Conclusion
The objective of this research is to investigate the relationship between an audit client’s workforce environment and audit
fees in an international setting. Specifically, we posit and find a better workforce environment may encourage employees to
dedicate themselves to their work, resulting in higher productivity and firm performance; as a result, this dedication
decreases client-specific risks and, thus, audit fees. In addition, our results also show that the negative association between
audit fees and the workforce environment is stronger in countries with a more flexible labor market, presumably because the
low hiring and firing costs in such markets encourage employees to be diligent and to cope with both external opportunities
and internal job loss threats, which will further decrease client-specific risks. We also posit that both firms’ financial report-
ing quality and media coverage of workforce controversies might mediate the effect of the workforce environment on audit
fees; however, we find supportive evidence for the media coverage channel only.
Our research is related to the literature studying the consequences of treating employees well from an external stake-
holder’s perspective (i.e., an auditor). Our study, in particular, should be of interest to corporate insiders (i.e., managers
and directors), as our results show that a firm’s efforts to improve its workforce environment will be recognized by both
internal (i.e., employees) and external stakeholders (e.g., auditors). In addition, our study responds to the call for additional
international-level research, so we may better understand how the variations in institutional factors influence existing find-
ings. In particular, we extend and complement Huang et al.’s (2017) results by showing that country-level labor market flex-
ibility moderates the negative association between audit fees and the workforce environment. Our findings also improve our
understanding of the channels through which a favorable workforce environment contributes to reduced audit fee.
Data availability
Data are available from the public sources cited in the text.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have
appeared to influence the work reported in this paper.
Acknowledgements
This paper is derived from Xuan Sean Sun’s Ph.D. dissertation at the School of Accountancy, Massey University, Auckland,
New Zealand. We acknowledge the valuable comments and suggestions received from Jeffrey Ng (editor) and an anonymous
reviewer. We are grateful to Andrew Jackson and David Lont for their constructive comments and suggestions on an earlier
version. We also thank Michael Bradbury, Steven Cahan, Charl de Villiers, Noor Houqe, Tom Scott, Roger Simnett, Jin Zhang,
Yuyu Zhang (AFAANZ discussant), the 2019 Eleventh Annual Quantitative Accounting Research Symposium and Consortium
participants at the University of Auckland, participants at the 2019 Accounting and Finance Association of Australia and New
Zealand annual conference, and seminar participants at Massey University for their useful comments.
Appendix A
Components of Workforce Environment Index
No. Description Positive or Negative Compliance

(%)
1 Does the company have a policy to drive diversity and equal opportunity? P: 1 for Yes, 0 for No 71.71%
2 Has the company set targets or objectives to be achieved on diversity and P: 1 for Yes, 0 for No 12.90%
equal opportunity?
3 Does the company claim to provide flexible working hours or working P: 1 for Yes, 0 for No 27.65%
hours that promote a work-life balance?
4 Does the company have a policy to improve employee health & safety P: 1 for Yes, 0 for No 75.71%
within the company and its supply chain?
5 Does the company have an employee health & safety team? P: 1 for Yes, 0 for No 47.78%
6 Does the company have health and safety management systems in place P: 1 for Yes, 0 for No 38.26%
like the OHSAS 18001 (Occupational Health & Safety Management
System)?
7 Does the company claim to provide day care services for its employees? P: 1 for Yes, 0 for No 13.24%
8 Does the company report on policies or programs on HIV/AIDS for the P: 1 for Yes, 0 for No 13.18%
workplace or beyond?
9 Does the company have a policy to improve the skills training of its P: 1 for Yes, 0 for No 62.86%
employees?
10 Does the company train its executives or key employees on health & P: 1 for Yes, 0 for No 62.89%
safety?
11 Does the company train its executives or key employees on employee P: 1 for Yes, 0 for No 9.22%
health & safety in the supply chain?
12 Does the company have a policy to improve the career development P: 1 for Yes, 0 for No 60.05%
paths of its employees?
13 Does the company claim to favor promotion from within? P: 1 for Yes, 0 for No 32.32%
14 Does the company claim to provide regular staff and business P: 1 for Yes, 0 for No 49.01%
management training for its managers?
15 Does the company provide training in environmental, social or P: 1 for Yes, 0 for No 15.13%
governance factors for its suppliers?
16 Does the company have a policy to support the skills training or career P: 1 for Yes, 0 for No 68.76%
development of its employees?
17 Is the company under the spotlight of the media because of a controversy N: 1 for No, 0 for Yes 96.69%
linked to the company’s employees, contractors or suppliers due to wage,
layoff disputes or working conditions?
18 Has there have been a strike or an industrial dispute that led to lost N: 1 for No, 0 for Yes 97.49%
working days?
19 Has an important executive management team member or a key team N: 1 for No, 0 for Yes 94.20%
member announced a voluntary departure (other than for retirement) or
has been ousted?
20 Total number of announced lay-offs by the company divided by the total N: 1 for no layoffs, 0 for 89.84%
number of employees. having lay-offs
N
Note: The qualitative measures provide binary information on whether a firm has a particular attribute, while the quanti-
tative measure gives information about actual lay-off activities. For qualitative measures with positive attributes, we assign
a score of 1 to a firm that has the attribute, and zero otherwise. For example, a firm will receive a score of 1 if it has an
employee health and safety team, and zero otherwise. For qualitative measures with negative attributes, we apply a score
of 1 to a firm that lacks the attribute, and zero otherwise. For instance, we give a score of 1 to a firm that has no strikes
or industrial disputes during the current year, and zero otherwise. For the quantitative measure, a firm receives a score of
1 if it does not lay off its employees, and zero otherwise.
Appendix B
Variable Definitions
Variable Definition Sources

Dependent variable
LNAF Natural logarithm of audit and audit-related fees. Thomson Reuters
(TR) Fundamentals
Independent variable
WEI Workforce environment index, measured by 20 firm-level qualitative and Authors’ calculation;
quantitative workforce-related indicators, which measures the overall TR ESG database
workforce environment for employees.
WS Thomson Reuters ESG workforce score, which measures ‘‘a company’s TR ESG database
effectiveness towards job satisfaction, a healthy and safe workplace,
maintaining diversity and equal opportunities, and development opportunities
for its workforce.”
WS_A4 The equal weighted workforce score (WS_A4) as the average of employment ASSET4; Authors’
quality (SOEQ), health and safety (SOHS), training and development (SOTD), and calculation
diversity and opportunities scores (SODO).
Firm-level control variable

SIZE Natural logarithm of total assets. TR Fundamentals
LEV The leverage ratio, defined as the sum of short- and long-term debt divided by TR Fundamentals
total assets.
INVREC The sum of inventories and receivables divided by total assets. TR Fundamentals
ROA Net income divided by total assets. TR Fundamentals
NBS Natural logarithm of the number of business segments. Worldscope
NGS Natural logarithm of the number of geographic segments. Worldscope
LOSS Dummy variable coded as 1 for firms reporting negative income before TR Fundamentals
extraordinary items, 0 otherwise.
SPECIAL Dummy variable, coded as 1 for firms reporting special items, 0 otherwise. TR Fundamentals
CROSS Dummy variable, coded as 1 for firms cross-listing in the US, 0 otherwise. Worldscope
MTB Market to book ratio, defined as the firm market value divided by common TR Fundamentals
shareholder equity.
CURRENT Current ratio, defined as current assets divided by current liabilities. TR Fundamentals
TURN Turnover ratio, defined as net sales divided by total assets. TR Fundamentals
INTS International operation, coded as 1 for firms that have at least 10 percent Worldscope
international sales to total sales, 0 otherwise.
ISSUE Dummy variable, coded as 1 when the following conditions apply: long-term Worldscope
debt increased by at least 20 percent, or the number of common shares
increased by at least 10 percent after controlling for stock splits; 0 otherwise.
CSR Corporate social responsibility score, which ranges from 0 to 100. TR ESG database
BIGN Dummy variable, coded as 1 for firms audited by one of the Big 4 firms (i.e., TR Eikon
PricewaterhouseCoopers, Deloitte Touche Tohmatsu, KPMG, and Ernst &
Young), 0 otherwise.
AO Audit opinion, coded as 1 for firms receiving qualified opinions for the current Worldscope
year, 0 otherwise.
BUSY Dummy variable, coded as 1 for firms for which the fiscal year-end comes Worldscope
during an auditor busy season, 0 otherwise.
Firm-level governance control variable

BSIZE The total number of board members at the end of the fiscal year. TR ESG database
BIND The percentage of independent members on the board of directors. TR ESG database
ACM Dummy variable, coded as 1 for firms with an audit committee, 0 otherwise. TR ESG database
ACMIND The percentage of independent members on the audit committee. TR ESG database
ACMEXP Dummy variable, coded as 1 for firms that have an audit committee with at TR ESG database
least three members and at least one financial expert, 0 otherwise.
(continued on next page)

Appendix B (continued)
Variable Definition Sources

CEODUAL Dummy variable, coded as 1 for firms where the CEO and chair of the board are TR ESG database
the same person, 0 otherwise.
Country-level control variable

LAW Legal origin. Dummy variable, coded as 1 for common law countries, 0 for code La Porta et al. (2008)
law countries.
LNGDP Natural logarithm of gross domestic product (GDP) per capita in U.S. dollars. International
Monetary Fund
FDI Foreign direct investment divided by GDP. World Bank
DISCL CIFAR disclosure developed by Center for International Financial Analysis and CIFAR
Research. CIFAR (1995) creates a country-specific index by rating the annual
reports of at least three firms in every country for the inclusion or omission of
85 specific items. These 85 items include specific disclosures in the following
seven categories: general information (8 items), income statement (11 items),
balance sheet (14 items), funds flow statement (5 items), accounting policy
disclosure (20 items), shareholder information (17 items) and other supple-
mentary information (10 items). Each country is given a score that ranges from
0 to 85, with higher scores indicating greater disclosure.
CORRUP Perceived corruption index, ranging from 0 to 10. For convenient interpretation, Transparency
we use 10 minus the actual values. Thus, the higher the index, the higher the International
perceived corruption level.
EPL Employment protection legislation index. It consists of three category scores: OECD website
individual dismissal of regular workers (EPR), additional costs of collective
dismissals (EPC), and regulation of temporary contracts (EPT). EPL is measured
as the equally weighted score of EPR, EPC, and EPT. We multiply EPL by 1 for
ease of interpretation.
EFW Labor market regulations including hiring regulations and minimum wage The Fraser Institute’s
(5Bi), hiring and firing regulations (5Bii), centralized collective bargaining Economic Freedom
(5Biii), hours regulations (5Biv), mandated cost of worker dismissal (5Bv), and of the World
conscription (5Bvi). We use the average of these six components to construct
the EFW index.
Mediation variable
|DAC| The absolute value of discretionary accruals (DAC), generated from the modified TR Fundamentals;
Jones model (Dechow et al., 1995). We estimate the model for all firms in the Authors’ calculation
same country and industry (using the GICS industry group code) with at least
eight observations in an industry-year-country pair using the following
h i
DSALESi;t DRECEIVABLEi;t
¼ c0 TAi;t1 þ c1 þ c2 TAi;t1 þ ei;t , (4)
ACC i;t 1 PPEi;t
equation: TAi;t1 TAi;t1
where ACC is the total accruals calculated as earnings before extraordinary
items and discontinued operations, minus operating cash flows. DSALES is the
change in sales revenue in year t; DRECEIVABLE is the change in accounts
receivable in year t; and PPE is the gross value of property, plant, and
equipment at the end of year t. All variables, including the intercept, are
deflated by lagged total assets. Non-discretionary accruals (NDAC) is the
predicted value from the above equation, with DAC representing the residuals.
CONTRO Media coverage of workforce controversies, measured as the natural logarithm TR ESG database
of the total number of controversies published in the media related to
workforce plus one. The total number of controversies published in the media
related to workforce includes ‘‘Number of controversies published in the media
linked to workforce diversity and opportunity (e.g., wages, promotion,
discrimination and harassment),” ‘‘Number of controversies published in the
media linked to the workforce health and safety,” and ‘‘Number of
controversies published in the media linked to the company’s relations with
employees or relating to wage or wage disputes.”
References
Addison, J.T., Teixeira, P., 2003. The economics of employment protection. J. Labor Res. 24 (1), 85–128.
Aguinis, H., Edwards, J.R., Bradley, K.J., 2017. Improving our understanding of moderation and mediation in strategic management research. Organ. Res.
Methods 20 (4), 665–685.
Akerlof, G.A., 1982. Labor contracts as partial gift exchange. Q. J. Econ. 97 (4), 543–569.
Bae, K.H., Kang, J.K., Wang, J., 2011. Employee treatment and firm leverage: a test of the stakeholder theory of capital structure. J. Financ. Econ. 100 (1), 130–
153.
Ball, R., 2016. Why we do international accounting research. J. Int. Account. Res. 15 (2), 1–6.
Banker, R.D., Byzalov, D., Chen, L.T., 2013. Employment protection legislation, adjustment costs and cross-country differences in cost behavior. J. Account.
Econ. 55 (1), 111–127.
Barney, J., 1991. Firm resources and sustained competitive advantage. J. Manage. 17 (1), 99–120.
Baron, R.M., Kenny, D.A., 1986. The moderator–mediator variable distinction in social psychological research: conceptual, strategic, and statistical
considerations. J. Pers. Soc. Psychol. 51 (6), 1173–1182.
Ben-Nasr, H., Ghouma, H., 2018. Employee welfare and stock price crash risk. J. Corp. Financ. 48, 700–725.
Bjuggren, C.M., 2018. Employment protection and labor productivity. J. Public Econ. 157, 138–157.
Bloom, N., Kretschmer, T., Van Reenen, J., 2011. Are family-friendly workplace practices a valuable firm resource?. Strateg. Manag. J. 32 (4), 343–367.
Bronson, S.N., Ghosh, A. Al, Hogan, C.E., 2017. Audit fee differential, audit effort, and litigation risk: an examination of ADR firms. Contemp. Account. Res. 34
(1), 83–117.
Carcello, J.V., Hermanson, D.R., Neal, T.L., Riley, R.A., 2002. Board characteristics and audit fees. Contemp. Account. Res. 19 (3), 365–384.
Carey, P., Liu, L., Qu, W., 2017. Voluntary corporate social responsibility reporting and financial statement auditing in China. J. Contemp. Account. Econ. 13
(3), 244–262.
Center for International Financial Analysis and Research (CIFAR), 1995. International accounting and auditing trends. CIFAR Publications Inc, Princeton.
Chen, C., Chen, Y., Hsu, P.H., Podolski, E.J., 2016a. Be nice to your innovators: employee treatment and corporate innovation performance. J. Corp. Financ. 39,
78–98.
Chen, T.K., Chen, Y.S., Yang, H.L., 2019. Employee treatment and its implications for bondholders. Eur. Financ. Manag. 25 (4), 1047–1079.
Chen, Y., Ge, R., Zolotoy, L., 2017. Do corporate pension plans affect audit pricing? J. Contemp. Account. Econ. 13 (3), 322–337.
Chen, L., Srinidhi, B., Tsang, A., Yu, W., 2016b. Audited financial reporting and voluntary disclosure of corporate social responsibility (CSR) reports. J. Manag.
Account. Res. 28 (2), 53–76.
Cheng, B., Ioannou, I., Serafeim, G., 2014. Corporate social responsibility and access to finance. Strateg. Manag. J. 35 (1), 1–23.
Cho, M., Ki, E., Kwon, S.Y., 2017. The effects of accruals quality on audit hours and audit fees. J. Account. Audit. Financ. 32 (3), 372–400.
Choi, J.H., Kim, J.B., Liu, X., Simunic, D.A., 2008. Audit pricing, legal liability regimes, and big 4 premiums: theory and cross-country evidence. Contemp.
Account. Res. 25 (1), 55–99.
Choi, J.H., Kim, J.B., Liu, X., Simunic, D.A., 2009. Cross-listing audit fee premiums: theory and evidence. Account. Rev. 84 (5), 1429–1463.
Chung, H., Kallapur, S., 2003. Client importance, nonaudit services, and abnormal accruals. Account. Rev. 78 (4), 931–955.
Dalal, R.S., Baysinger, M., Brummel, B.J., Lebreton, J.M., 2012. The relative importance of employee engagement, other job attitudes, and trait affect as
predictors of job performance. J. Appl. Soc. Psychol. 42, E295–E325.
Dechow, P., Ge, W., Schrand, C., 2010. Understanding earnings quality: a review of the proxies, their determinants and their consequences. J. Account. Econ.
50 (2–3), 344–401.
Dechow, P.M., Sloan, R.G., Sweeney, A.P., 1995. Detecting earnings management. Account. Rev. 70 (2), 193–225.
Duellman, S., Hurwitz, H., Sun, Y., 2015. Managerial overconfidence and audit fees. J. Contemp. Account. Econ. 11 (2), 148–165.
Dyck, A., Morse, A., Zingales, L., 2010. Who blows the whistle on corporate fraud?. J. Finance 65 (6), 2213–2253.
Edmans, A., 2011. Does the stock market fully value intangibles? Employee satisfaction and equity prices. J. Financ. Econ. 101 (3), 621–640.
Edmans, A., 2012. The link between job satisfaction and firm value, with implications for corporate social responsibility. Acad. Manag. Perspect. 26 (4), 1–19.
Edmans, A., Li, L., Zhang, C., 2017. Employee Satisfaction, Labor Market Flexibility, and Stock Returns Around the World. Working paper. Business School,
London.
Faleye, O., Trahan, E.A., 2011. Labor-friendly corporate practices: is what is good for employees good for shareholders?. J. Bus. Ethics 101 (1), 1–27.
Fauver, L., McDonald, M.B., Taboada, A.G., 2018. Does it pay to treat employees well? International evidence on the value of employee-friendly culture. J.
Corp. Financ. 50, 84–108.
Fehr, E., Gächter, S., 2000. Fairness and retaliation: the economics of reciprocity. J. Econ. Perspect. 14 (3), 159–181.
Feldmann, H., 2009. The unemployment effects of labor regulation around the world. J. Comp. Econ. 37 (1), 76–90.
Filbeck, G., Preece, D., 2003. Fortune’s best 100 companies to work for in America: do they work for shareholders?. J. Bus. Financ. Account. 30 (5–6), 771–
797.
Fulmer, I.S., Gerhart, B., Scott, K.S., 2003. Are the 100 best better? An empirical investigation of the relationship between being a ‘‘great place to work” and
firm performance. Pers. Psychol. 56 (4), 965–993.
Gangl, M., 2003. The only way is up? Employment protection and job mobility among recent entrants to European labour markets. Eur. Sociol. Rev. 19 (5),
429–449.
Gong, S.X., Gul, F.A., Shan, L., 2018. Do auditors respond to media coverage? Evidence from China. Account. Horizons 32 (3), 169–194.
Griffin, P.A., Lont, D.H., Sun, E., 2008. Corporate governance and audit fees: evidence of countervailing relations. J. Contemp. Account. Econ. 4 (1), 18–49.
Guo, J., Huang, P., Zhang, Y., Zhou, N., 2016. The effect of employee treatment policies on internal control weaknesses and financial restatements. Account.
Rev. 91 (4), 1167–1194.
Gupta, K., Krishnamurti, C., 2018. Do countries matter more in determining the relationship between employee welfare and financial performance?. Int. Rev.
Financ. (in press).
Haltiwanger, J., Scarpetta, S., Schweiger, H., 2014. Cross country differences in job reallocation: the role of industry, firm size and regulations. Labour Econ.
26, 11–25.
Hansen, L.P., 1982. Large sample properties of generalized method of moments estimators. Econom. J. Econom. Soc. 50 (4), 1029–1054.
Hawn, O., Kang, H.G., 2018. The effect of market and nonmarket competition on firm and industry corporate social responsibility. Adv. Strateg. Manag. 38,
313–337.
Hay, D.C., Knechel, W.R., Wong, N., 2006. Audit fees: a meta-analysis of the effect of supply and demand attributes. Contemp. Account. Res. 23 (1), 141–191.
Hay, D., Knechel, W.R., Ling, H., 2008. Evidence on the impact of internal control and corporate governance on audit fees. Int. J. Audit. 12 (1), 9–24.
Hayes, A.F., 2009. Beyond Baron and Kenny: statistical mediation analysis in the new millennium. Commun. Monogr. 76 (4), 408–420.
Hayes, A.F., 2018. Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-based Approach. The Guilford Press, New York.
Hom, P.W., Kinicki, A.J., 2001. Toward a greater understanding of how dissatisfaction drives employee turnover. Acad. Manag. J. 44 (5), 975–987.
Hom, P.W., Lee, T.W., Shaw, J.D., Hausknecht, J.P., 2017. One hundred years of employee turnover theory and research. J. Appl. Psychol. 102 (3), 530–545.
Houston, R.W., Peters, M.F., Pratt, J.H., 2005. Nonlitigation risk and pricing audit services. Audit. J. Pract. Theory 24 (1), 37–53.
Huang, M., Masli, A., Meschke, F., Guthrie, J.P., 2017. Clients’ workplace environment and corporate audits. Audit. J. Pract. Theory 36 (4), 89–113.
Huselid, M.A., 1995. The impact of human resource management practices on turnover, productivity, and corporate financial performance. Acad. Manag. J.
38 (3), 635–672.
Ichino, A., Riphahn, R.T., 2005. The effect of employment protection on worker effort: absenteeism during and after probation. J. Eur. Econ. Assoc. 3 (1), 120–
143.
Ioannou, I., Serafeim, G., 2012. What drives corporate social performance? The role of nation-level institutions. J. Int. Bus. Stud. 43 (9), 834–864.
Ioannou, I., Serafeim, G., 2017. The Consequences of Mandatory Corporate Sustainability Reporting. Working paper. Harvard Business School.
Jaggi, B., Low, P.Y., 2011. Joint effect of investor protection and securities regulations on audit fees. Int. J. Account. 46 (3), 241–270.
Ji, X.D., Lu, W., Qu, W., 2018. Internal control risk and audit fees: evidence from China. J. Contemp. Account. Econ. 14 (3), 266–287.
Karpoff, J.M., Lott, J.R., 1999. On the determinants and importance of punitive damage awards. J. Law Econ. 42 (S1), 527–573.
Kim, J.B., Liu, X., Zheng, L., 2012. The impact of mandatory IFRS adoption on audit fees: theory and evidence. Account. Rev. 87 (6), 2061–2094.
Kleibergen, F., Paap, R., 2006. Generalized reduced rank tests using the singular value decomposition. J. Econom. 133 (1), 97–126.
Kothari, S.P., Leone, A.J., Wasley, C.E., 2005. Performance matched discretionary accrual measures. J. Account. Econ. 39 (1), 163–197.
Kuo, N.T., Lee, C.F., 2016. A potential benefit of increasing book–tax conformity: evidence from the reduction in audit fees. Rev. Account. Stud. 21 (4), 1287–
1326.
La Porta, R., Lopez-de-Silanes, F., Shleifer, A., 2008. The economic consequences of legal origins. J. Econ. Lit. 46 (2), 285–332.
Larcker, D.F., Rusticus, T.O., 2010. On the use of instrumental variables in accounting research. J. Account. Econ. 49 (3), 186–205.
Lee, J., Kim, H., 2016. Do employee relation responsibility and culture matter for firm value? International evidence. Pacific Basin Financ. J. 40, 191–209.
Levine, D.I., 1992. Can wage increases pay for themselves? Tests with a productive function. Econ. J. 102 (414), 1102–1115.
LópezPuertas-Lamy, M., Desender, K., Epure, M., 2017. Corporate social responsibility and the assessment by auditors of the risk of material misstatement. J.
Bus. Financ. Account. 44 (9–10), 1276–1314.
Marquaridt, D.W., 1970. Generalized inverses, ridge regression, biased linear estimation, and nonlinear estimation. Technometrics 12 (3), 591–612.
O’Keefe, T.B., Simunic, D.A., Stein, M.T., 1994. The production of audit services: evidence from a major public accounting firm. J. Account. Res. 32 (2), 241–
261.
Ormazábal, G., 2018. The role of stakeholders in corporate governance: a view from accounting research. Found. TrendsÒ Account. 11 (4), 193–290.
Pagano, M., Volpin, P.F., 2005. The political economy of corporate governance. Am. Econ. Rev. 95 (4), 1005–1030.
Pfeffer, J., Villeneuve, F., 1994. Competitive advantage through people: unleashing the power of the work force, vol. 61. Harvard Business School Press
Boston, MA.
Pratt, J., Stice, J.D., 1994. The effects of client characteristics on auditor litigation risk judgments, required audit evidence, and recommended audit fees.
Account. Rev. 69 (4), 639–656.
Preacher, K.J., Hayes, A.F., 2004. SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behav. Res. Methods, Instruments,
Comput. 36 (4), 717–731.
Redmayne, N.B., Bradbury, M.E., Cahan, S.F., 2010. The effect of political visibility on audit effort and audit pricing. Account. Financ. 50 (4), 921–939.
Rothschild, J., Miethe, T.D., 1999. Whistle-blower disclosures and management retaliation: the battle to control information about organization corruption.
Work Occup. 26 (1), 107–128.
Roychowdhury, S., 2006. Earnings management through real activities manipulation. J. Account. Econ. 42 (3), 335–370.
Shapiro, C., Stiglitz, J.E., 1984. Equilibrium unemployment as a worker discipline device. Am. Econ. Rev. 74 (3), 433–444.
Simunic, D.A., 1980. The pricing of audit services: theory and evidence. J. Account. Res. 18 (1), 161–190.
Simunic, D.A., 1984. Auditing, consulting, and auditor independence. J. Account. Res. 22 (2), 679–702.
Simunic, D.A., Stein, M.T., 1996. The impact of litigation risk on audit pricing: a review of the economics and the evidence. Audit. J. Pract. Theory 15 (Suppl.),
119–134.
Taylor, F.W., 1914. The Principles of Scientific Management. Harper Brothers, New York.
Taylor, M.H., Simon, D.T., 1999. Determinants of audit fees: the importance of litigation, disclosure, and regulatory burdens in audit engagements in 20
countries. Int. J. Account. 34 (3), 375–388.
Thomson Reuters, 2018. Thomson Reuters ESG scores. https://financial.thomsonreuters. com/content/dam/openweb/documents/pdf/financial/esg-scores-
methodology.pdf (accessed 15 May 2018).
Turban, D.B., Cable, D.M., 2003. Firm reputation and applicant pool characteristics. J. Organ. Behav. 24 (6), 733–751.
Wright, P.M., McMahan, G.C., McWilliams, A., 1994. Human resources and sustained competitive advantage: a resource-based perspective. Int. J. Hum.
Resour. Manag. 5 (2), 301–326.
Zaman, M., Hudaib, M., Haniffa, R., 2011. Corporate governance quality, audit fees and non-audit services fees. J. Bus. Financ. Account. 38 (1–2), 165–197.
Zhao, X., Lynch, J.G., Chen, Q., 2010. Reconsidering Baron and Kenny: myths and truths about mediation analysis. J. Consum. Res. 37 (2), 197–206.

Workforce Environment and Audit Fees (2020)

Uploaded by

Copyright:

Available Formats

Workforce Environment and Audit Fees (2020)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Workforce Environment and Audit Fees (2020)

Uploaded by

Copyright:

Available Formats

Journal of Contemporary Accounting and Economics 16 (2020) 100182

Contents lists available at ScienceDirect

Workforce environment and audit fees: International evidence

2. Literature review and hypotheses development

3.1. Model specification

3.2. Measurement of audit fees

3.3. Measure of workforce environment index

3.4. Measurement of labor market flexibility

Industry group Name N % of N

4.1. Descriptive statistics

Panel A: Firm-level variables

Note: Variables are defined in Appendix B.

4.2. Correlation analysis

4.4. Workforce environment and audit fees: moderation results for H2

(1) (2) (3) (4)

Note: Robust t-statistics are in brackets. Variables are defined in Appendix B.

Note: Robust t-statistics are in brackets. Variables are defined in Appendix B.

4.5. Mediation test results

Mi;t ¼ b0 þ b1 WEIi;t þ Control v ariables þ Fixed Effects þ ei;t ; ð3BÞ

5. Additional robustness tests

5.1. Alternative measure of workforce environment

5.2. Exclusion of multinational firms

5.3. Control for non-audit services fees

5.4. Additional analyses

Panel A: Alternative independent variables

Panel B: Non-US results and results with lagged independent variables

5.5. Endogeneity test

5.6. Alternative argument on workforce environment

(1) (2) (3)

Note: Robust t-statistics are in brackets. Variables are defined in Appendix B.

Declaration of Competing Interest

Components of Workforce Environment Index

No. Description Positive or Negative Compliance

Variable Definition Sources

Firm-level control variable

Firm-level governance control variable

(continued on next page)

Variable Definition Sources

Country-level control variable

You might also like