Introduction Big 4 Office Personnel and Audit Quality

CHAPTER 1

INTRODUCTION

Since the founding of the Public Company Accounting Oversight Board (PCAOB)

in 2002, audit firm inspection reports have shown that audit quality is inconsistent at the

firm level, office level, and engagement team level (PCAOB, 2015). In light of these

inconsistencies, the PCAOB established a project in 2013 involving regulators,

professionals, and academics to “identify a portfolio of quantitative measures of public

company auditing (called indicators), whose consistent use may enhance dialogue about

and understanding of audits and ways to evaluate their quality; and to explore how and by

whom the portfolio of indicators can best be used” (PCAOB, 2015, 2). The resulting

portfolio identified 28 potential audit quality indicators (AQIs), grouped by audit

professionals, audit processes, and audit results (See Table 1). The

PCAOB’s feedback from regulators, professionals, and academics showed that key areas

for future discussion on audit quality should include experience (AQI 6), workload (AQI

2-3), and human capital (AQI 1, 4-5, 8), all of which focus on the professionals

conducting the audit engagement.

The PCAOB places a high level of scrutiny on audit engagement quality to ensure

that users have the best possible financial information with which to make better and more

confident decisions (Fellner & Maciejovsky, 2007; Nelson, 1970; Akerlof, 1970).

However, the PCAOB commonly focuses on auditor reporting, client characteristics, and

audit firm internal quality control systems in its evaluation of audit quality, rather than the

characteristics and backgrounds of the audit professionals at the firm.

TABLE 1

PCAOB Audit Quality Indicators (PCAOB, 2015, p. 13)

1

2

3

4

5

6

7

8

9

10

Staffing Leverage

Partner Workload

Manager and Staff Workload

Technical Accounting and Auditing Resources

Persons with Specialized Skill and Knowledge

Experience of Audit Personnel

Industry Expertise of Audit Personnel

Turnover of Audit Personnel

Amount of Audit Work Centralized at Service Centers

Training Hours per Audit Professional

11

Audit hours and Risk Areas

12

Allocation of Audit Hours to Phases of the Audit

13

14

15

16

17

18

19

Results of Independent Survey of Firm Personnel

Quality Ratings and Compensation

Audit Fees, Effort, and Client Risk

Compliance with Independence Requirements

Investment in Infrastructure Supporting Quality Auditing

Audit Firms' Internal Quality Review Results

PCAOB Inspection Results

20

Technical Competency Testing

21

22

23

24

25

26

27

Frequency and Impact of Financial Statement Restatements for Errors

Fraud and Financial Reporting Misconduct

Inferring Audit Quality from Measures of Financial Reporting Quality

Timely Reporting of Internal Control Weaknesses

Timely Reporting of Going Concern Issues

Results of Independent Surveys of Audit Committee Members

Trends in PCAOB and SER Enforcement Proceedings

28

Trends in Private Litigation

U.S. studies often attempt to measure audit quality using Big 4 office fees, client

characteristics including financial reporting and audit outcomes, and audit engagement

problems (Knechel, Pevzner, Shefchik, & Velury, 2013; DeFond & Zhang, 2014).

Knechel et al. (2013), in addition to the PCAOB, also called for extended research

on audit quality by uncovering new sources of data inputs for audit department personnel

that might better evaluate why audit quality is inconsistent. Historically, archival studies

have had difficulty in peeling back the layers of the Big 4 office to delve further into its

individual auditor characteristics as information about audit department personnel is not

publicly disclosed by Big 4 firms in the U.S (Knechel et al., 2013; Francis, 2011).

Utilizing a uniquely compiled data source, the overarching purpose of this study is to

investigate the following question: what is the impact of Big 4 office auditor human

capital, auditor experience, and auditor workload compression on audit quality? This

study addresses the question in a fashion that heretofore has been limited in the U.S.

Rather than solely using external office characteristics such as the number of

clients or total fees, this study includes audit department personnel inputs. Typically, Big

4 office personnel have been examined using experimental methods or by using available

partner information in non-U.S. countries (Cahan & Sun, 2015; Gul, Wu, & Yang, 2013).

While these studies provide some insights, they are often limited to either partner level

analysis or the internal validity of experimental studies trades-off for external validity.

That is, from these studies, we are unable to draw implications about how audit

department level personnel influence audit quality. Therefore, using established measures

of audit quality that are discussed in detail in Chapter 3, and a unique dataset, Big 4 office

personnel are examined at the audit department level, described below.

Audit Human Capital. Human capital refers to the human resources available to

each Big 4 office. This study will examine the relationship between audit human capital

available at each office and audit quality, based on the total audit department headcount

compared to the total office headcount; the headcount of partners, principals, senior

managers, managers, senior associates, and associates compared to the total audit

department headcount; and the influence of education on the quantity and quality of

partners, principals, senior managers, managers, senior associates, and associates available

in each audit department.

Experience. Experience refers to the personal tenure of each Big 4 professional.

This study will examine the relationship between audit personnel experience in each

office and audit quality, based on audit partner, principal, senior manager, manager, senior

associate, and associate experience in years at the audit office; at the audit firm; and as an

auditor regardless of the audit firm.

Workload. Workload refers to the volume of audit engagements at each Big 4

office. This study will examine the relationship between auditor workload and audit

quality, based on the number of audit clients compared to the total audit department

headcount for each office, and the fiscal year-end groupings of audit clients compared to

the headcount of audit partners, principals, senior managers, managers, senior associates,

and associates for each office.

The models for this study include numerous control variables to mitigate the

indirect effects of client engagement and auditor risk factors. The study is limited by the

data source for audit department-level auditor characteristics. The Big 4 audit

department-level data comes from LinkedIn (using a Premium version which provides

advanced search options and more complete profile information) with profiles populated

by individual employees of the Big 4. The assumption is made that individuals placing

their information on LinkedIn have incentives to ensure the information is accurate and

credible as the hosting organization provides networks and career opportunities to both

employees and employers, and serves as a reference check site for human resource

managers (Brown & Vaughn, 2011; Zide, Elman, & Shahani-Denning, 2014). In addition,

the Big 4 encourages their current employees and alumni to establish and maintain

LinkedIn profiles connected to the official LinkedIn Big 4 groups, knowing that these

individuals often go on to achieve influential positions at companies that have the

potential to become future clients2. To provide some validity to the data extracted, the

LinkedIn personnel data is compared to data provided by a Big 4 office, Form AP data,

and headcount information listed on a Big 4 firm website. The study is also limited by the

sample year as the data provided in LinkedIn is at a point in time and the data is extracted

for this study in 2018.

This study adds to the body of literature analyzing Big 4 audit quality. Academic

research continues to present a tension surrounding Big 4 audit procedures and

inconsistent audit quality among audit offices in each Big 4 firm (Reynolds & Francis,

2000; Choi, Kim, Kim & Zang, 2010). The PCAOB requires registered audit firms to

maintain a system of quality controls relative to the audit function, to ensure that proper

professional and ethical standards are being followed (PCAOB CQ Section 20), and to

ensure that firm practices, procedures, and reporting standards are consistently applied at

each audit office in accordance with firm audit methodology (Our System of Audit

Quality Controls, 2015; Our Commitment to Audit Quality, 2017; Our Focus on Audit

Quality, 2017; U.S. Audit Quality Report, 2016). Yet, academic research findings and

PCAOB inspections indicate that firm practices, procedures, and reporting standards are

not consistently applied to all offices due to the sharing of information and experiences

among peers, and because organizational leaders will develop personal best practices that

may differ by location (Diaz, Martin & Thomas, 2017; Francis & Yu, 2009; Francis,

Michas, & Yu, 2013; Hambrick & Mason, 1984). The result is that higher audit quality is

purported to occur at offices with the most audit human capital and/or the most auditor

experience. This study extends the academic research on audit quality by being among

the first to examine in detail the audit human capital characteristics at the Big 4 office

level and audit staff personnel level.

This study makes a significant contribution to the profession and to regulators by

providing information about audit staff that can form audit engagement teams. The Wall

Street Journal recently expressed concerns that “high-profile cases keep emerging in

which significant issues or fraud get past an auditor” (Rapoport, 2018). The PCAOB firm

inspection reports support these concerns, showing that uneven audit quality is occurring

among audit firms; offices within the same firm; and engagement teams within the same

office. Past-PCAOB member Steven Harris recently stated that he believed uneven audit

quality occurred largely because of the individuals performing the audit due to “poor

supervision, failure to exercise appropriate professional skepticism, ignoring contradictory

evidence, poor audit planning, lack of training or knowledge of audit personnel, and tight

deadlines” (Harris, 2016). In public practice, audit quality is typically measured directly

through annual external inspections (PCAOB, 2004). In these inspections, members of

the PCAOB organization examine the workpapers, reports, and methodologies of the firm

to determine if the audit was conducted properly in accordance with the applicable

auditing standards. While the overall findings of the inspections are made public, certain

information regarding the specific firm office is not released. The PCAOB has indicated

that they have incorporated the 28 AQIs into their programs to make their inspections

more robust (PCAOB, 2017). The results of this study may be helpful to the PCAOB as it

considers its recommendations for enhancing the quality of the audit.

The remainder of this paper is organized as follows: Chapter 2 discusses prior

relevant research and develops the hypotheses. Chapter 3 presents the research design,

sample and data sources, and discusses the proposed analyses. Chapter 4 discusses data

analysis and presents the results. Chapter 5 provides a summary and conclusion, and

identifies limitations and opportunities for future research.

CHAPTER 2

LITERATURE REVIEW AND HYPOTHESES

The Importance of Audit Quality

In 1976, Jensen and Meckling expanded on management/stockholder information

asymmetry with their seminal theory of the firm. According to their theory, management,

as agents, is presumed to act on behalf of shareholders, the principals (Jensen &

Meckling, 1976). Although shareholders have an expectation that management will make

strategic decisions that increase value, generate cash flows, and manage operations in

accordance with company policy, management may act in its own best interest to achieve

personal goals. Users are often unaware of management’s self-serving behavior due to

information asymmetry. This problem is called agency cost and can be mitigated by

intermediation of external auditors, who are often paid significant fees to provide high-

quality audits of the financial statements.

Audit quality has been historically defined as both the ability of an auditor to

discover financial reporting issues and the willingness to report them (DeAngelo, 1981).

As the body of literature surrounding audit quality has grown, researchers have

developed numerous ways to measure external audit quality to varying degrees. To more

precisely organize the myriad characteristics that define audit quality, Francis (2011)

developed a framework of analysis units, consisting of audit inputs, audit processes,

accounting firms, guidance, audit industries and markets, institutions, and economic

9

8

consequences of audit outcomes. Knechel et al. (2013) extended Francis’s (2011)

primarily archival synthesis to include behavioral, experimental, and survey research,

while further developing the concepts by which inputs, processes, and outcomes indicate

audit quality. Both studies note that the most difficult area to study is audit inputs since

information about the audit engagement team planning and procedures, audit department

composition, and individual auditor experience have historically been difficult to gather

due to its personal or proprietary nature.

Because the nature of auditing necessitates individual judgments, it is natural that

audit quality is highly dependent on the inputs of individuals performing the audit. As

Knechel et al. (2013, 391) state: “The ability to make sound judgments directly

influences the quality of the audit, so the better the personnel, the better the outcome of

the audit is likely to be.” Auditor judgments are often influenced by the conflicting

pressures of client-desired outcomes and potential regulatory or reputation risks (Haynes,

Jenkins, & Nutt, 1998), although auditors with greater overall and industry-specific

experience are better able to withstand client pressures and thus make higher quality

decisions (Frederick & Libby, 1986; Beck & Wu, 2006). However, the firms themselves,

and offices within the firms, must be able to find the appropriate individuals in order to

mollify the audit quality deterioration that can occur due to budget overruns and

busyseason time compressions (DeZoort & Lord, 1997; McDaniel, 1990). Thus, the

ability of firms to achieve high audit quality is directly influenced by their ability to

accumulate the right amount and composition of human capital.

10

Audit Human Capital and Audit Quality

The theory of human capital suggests that skills and knowledge have value to a

firm and that this valuable investment is made due to a conscious effort from the firm

(Schultz, 1961). Human capital can be acquired and/or it can be developed internally.

The majority of Big 4 human capital acquisitions begin at the staff level, so that they can

be “molded” in accordance with firm culture (Hermanson, Houston, Stefaniak, &

Wilkins, 2016). Once human capital is acquired by a firm, the development process

typically consists of three components. First, auditors are given formal professional

education regarding firm methodology, audit standards, and accounting standards (Yang,

Chen, & Yang, 2013; Schultz, 1961; Lepak & Snell, 1999; Fung, Gul & Krishnan, 2012).

Second, auditors are developed through on-the-job training, by engagement experience,

observation of other auditors, and monitoring with feedback from supervisors (Schultz,

1961; Neumann, 1981). Third, auditors increase their skills and knowledge by taking

advantage of specialization and/or office rotation opportunities, which provide auditors

with unique skill sets and industry expertise that can increase value to clients and reduce

costs to the firm (Schultz, 1961; Fung et al., 2012).

Once a firm acquires and develops its audit human capital to a minimum level, it

depends on these resources in order to achieve and maintain high audit quality, and

continues in-house training and education (Pfiefier & Salancik, 1978; Hillman & Dalziel,

2003; Francis & Yu, 2009). Prior literature typically assumes that the largest offices will

have the most audit human capital, and therefore, will provide higher audit quality.

DeAngelo (1981) hypothesized that Big 4 audit firms produce higher quality audits than

non-Big 4 firms due to more available resources and technology. The findings of

11

Reynolds and Francis (2000) support DeAngelo’s proposition. Similarly, using the

number of clients to proxy for the size of human capital, Choi, Kim, Kim, and Zang

(2010) find the Big 4 offices with more clients produce high-quality audits. Francis and

Yu (2009) propose that larger offices presumably with more in-house audit personnel

experiences provide greater amounts of audit human capital than smaller offices due to

the decentralized nature of the Big 4 office structure. Further, Francis and Michas (2013)

assume that offices with the most in-house audit personnel experiences are able to apply

their industry-specific knowledge and skills to client engagements, resulting in high audit

quality. Therefore, this study presents the following hypothesis:

H1a: There is a positive relationship between audit human capital available at the

office level and audit quality.

Staffing leverage, a PCAOB audit quality indicator, refers to the availability of

senior-level personnel relative to the amount of requisite work for which they provide

oversight (PCAOB 2015). When audit firms have sufficient individuals with valuable

skills and knowledge (i.e. human capital), then they are better placed to provide

highquality audits. Management literature presents numerous studies discussing the

importance of senior-level leadership on company performance and strategic

decisionmaking (for example Herzberg, 1987 and Vera & Crossnan, 2004). Similarly,

publications from regulatory bodies such as the PCAOB have stated that “tone” which

exists at the senior-levels of an organization is the key element that “drives a firm’s

culture and personnel management” (PCAOB, 2015, 13). This observation exists in the

extant audit quality literature as well. Using Chinese data, both Cahan and Sun (2015)

and Gul et al. (2013) examined the individual effects of the signing audit partner on audit

12

quality, and find that experience, both with the firm and overall, is positively associated

with audit quality, measured by more opinion modifications and fewer financial

restatements. However, for high audit quality to persist, firms must be able to acquire

sufficient experienced individuals. Indeed, Hossain et al., (2017), using Japanese data,

examined the composition of CPAs, Junior CPAs, and other professional staff (classified

in accordance with Japanese financial disclosure requirements) on audit teams and found

that audit teams with greater numbers of CPAs produce higher quality audits, when

measured by the propensity to issue going concern opinion modifications and

discretionary accruals. Similarly, Francis and Yu (2009) show that U.S. offices with more

clients (which they assume indicates a higher number of audit professionals) are more

likely to issue going concern opinion modifications and less likely to have greater

discretionary accruals. Together, these studies indicate that audit human capital for senior

level auditors will affect audit quality to a greater extent than audit human capital for

lower level auditors for three primary reasons. First, while lower level auditors tend to

perform the vast majority of fieldwork, their tasks are typically highly supervised and

non-judgmental. Second, senior level auditors tend to make judgments and decisions that

more pointedly affect audit quality, such as the elected omission of an audit adjustment or

a negotiation with the audit committee regarding financial statement materiality (Ng &

Tan, 2003). And finally, clients that possess greater risks of financial misstatements

demand more attention from senior level auditors in order to achieve a higher level of

audit quality (O’Keefe, Simunic, & Stein, 1994; Stein, Simunic, & O’Keefe, 1994).

Therefore, this study presents the following hypothesis:

13

H1b: The positive relationship between audit human capital available at the office

level and audit quality is more pronounced in audit offices that possess more senior-level

audit human capital.

Because accounting employee turnover rates have ranged from 10%-15% as

recently as 2017 (according to insidepublicaccounting.com, an organization that annually

surveys hundreds of accounting firms), it is vital that auditing departments acquire

sufficient strong-foundational human capital on which to build experience over time.

Prior studies show that elite and specialized colleges and universities prepare their

accounting students for entry level positions better than others. For instance, Hardin and

Stocks (1995) find that recruiters view accounting students from an accounting program

separately accredited by the American Assembly of Collegiate Schools of Business

(AACSB) as more attractive hires than students not from an accounting program

separately accredited by the AACSB. Also, Bunker and Harris (2014) find that

accounting students attending an AACSB college or university perform better on the CPA

exam than students attending a non-AACSB college or university. As auditors gain more

experience on the job, however, it is likely that any educational effects on audit quality

will be lessened. For example, although senior level professionals might possess Ivy-

League education or come from AACSB or flagship universities (Erkens & Bonner,

2013; Badolato, Donelson, & Ege, 2014), other factors like performance and relationship

development also play a vital role in achieving senior level status (Westphal & Stern,

2006). Therefore, this study presents the following hypothesis:

14

H1c: The positive relationship between education quality possessed by audit

human capital available at the office level and audit quality decreases with the level of

seniority of audit staff.

Audit Personnel Experience and Audit Quality

Upper echelon theory proposes that the top management teams will make

strategic decisions based on individual interpretations and that these interpretations tend

to be formed by personal experiences and characteristics (Hambrick & Mason, 1984;

Hambrick, 2007). In an audit firm, the more authoritative individuals (e.g. partners,

principals, and managers) may develop internal best practices based on their experiences

which will influence organizational outcomes, such as the strategic selection of a client;

implementation of a procedure; or the effectiveness of an accounting judgment. Because

of these experiential effects, separate offices in the same accounting firm could have

different accounting valuations, classifications, audit judgments, and financial disclosures

leading to inconsistent audit quality across the firm.

Big 4 firms’ quality controls and audit methodologies attempt to mitigate

inconsistent audit quality by transferring knowledge and experiences across the firm

offices (for example: through database sharing and internal training) to achieve

consistently high audit quality (Our System of Audit Quality Controls, 2015; Our

Commitment to Audit Quality, 2017; Our Focus on Audit Quality, 2017; U.S. Audit

Quality Report, 2016). Further, in certain complex situations defined by the firm,

decisions are required to be made at the national office to maintain consistency in

judgments. Yet, despite these efforts, research proposes that high audit quality is

achieved by the largest, most experienced, highest billing offices in the firm rather than

15

consistently throughout all offices (Francis & Yu, 2009; Francis et al., 2013). This

inconsistency appears largely due to the inability of audit firms to successfully transfer

knowledge and experiences among offices (Bamber, Jiang, & Wang, 2010; Francis et al.,

2013).

Due to the difficulties of transferring experienced-based knowledge among all

offices in the audit firm, it stands to reason certain offices might perform better than

others, because the senior-level leadership at each Big 4 office is likely to assemble a

personnel composition that will lead to the best possible quality based on their personal

experiences and knowledge. Audit teams are commonly comprised of partners,

principals, senior managers, managers, senior associates, and associates. Additionally,

each audit team is comprised of individuals with varying degrees and types of personal

experience. Because of these differences and because the audit team authoritative

individuals may utilize personal interpretations, audit quality may vary among the offices

in the same Big 4 firm. The PCAOB (with the support of numerous audit firms)

suggested that information concerning auditor experience could provide vital insight into

the mechanics of the external audit (PCAOB, 2015). Researchers also suggest that

examinations of the personnel comprising the audit teams at each individual office could

provide a more precise measure of audit quality (Knechel et al., 2013), yet information

regarding these inputs and processes has been difficult to gather from an archival

perspective, as U.S. Big 4 firms do not publicize this type of data.

Several studies have performed behavioral research using experience levels to

investigate decision-making. Simon and Chase (1973) used a chess simulation game to

evaluate how chess players of varying skill level made decisions based on their ability to

recall certain positions and potential future moves. The authors found evidence that

16

master chess players had a greater recall ability than novice chess players because

strategy is a component of the placement of chess pieces. Similarly, Moeckel (1990)

found that more experienced auditors were able to incorporate recall and reconstruction

into their decision-making more effectively than auditors with less experience. Moeckel

also found that, although the auditors made errors in the experiment, the errors tended to

occur less often with more experienced auditors. Naturally, as auditors acquire more

experience both through client engagements and through training, their error rate

decreases. This occurs in part due to the awareness of a larger population of errors that

could potentially happen and because of their client experience and training in specific

transactions cycles (Libby & Frederick, 1990).

Using experiments, Kinney and Uecker (1982) and McDaniel and Kinney (1995)

analyzed audit senior associates and their abilities to make analytical decisions. They

found that senior associates rarely stray from the use of decision aids, methodology, and

prior year information to make audit decisions. Similarly, other behavioral studies show

that although auditors’ risk assessment approach varies by firm, staff auditors tend to rely

primarily on firm methodology throughout the audit engagement process (Jiambalvo &

Waller, 1984; Wilks & Zimbelman, 2004; Zimbelman, 1997). Alternatively, Cianci et al.

(2017) found that in certain situations, partners relied on personal judgments rather than

firm methodology or other regulations. These results indicate that more experienced

auditors (e.g., partners) should have a greater individualized impact on the audit

engagement than auditors with less experience (e.g., senior associates and associates).

A limited amount of archival research has examined the impact of auditor

experience on audit quality. Using available Japanese data, Hossain et al., (2017) found

some evidence to indicate a positive association between senior and staff auditors’

17

experience and audit fees, but not with any other measures of audit quality. Other

archival studies, however, have much stronger findings regarding audit quality when

using measurements at the audit partner level. Cahan and Sun (2015) utilized upper

echelon theory to determine the impact of Chinese audit partner experience on audit fees

and audit quality. In China, personal information about audit partners is more readily

available than the U.S., as audit partners are required to sign the audit report and disclose

experience to the Chinese Securities Regulatory Commission. Using audit fees and

discretionary accruals as proxies for audit quality, this study found that partners with

more years of experience provide higher quality non-U.S. audits than partners with less

experience. Gul, Wu, and Yang (2013) also used Chinese data to examine audit partners

that signed off on the audit reports. They too found a significant relationship between

audit partner experience and audit quality. Therefore, the following hypothesis is

proposed:

H2a: - The positive relationship between audit personnel experience and audit

quality is greater at more senior levels than at less senior levels.

While there is a general positive relationship between experience and audit

quality, too much experience could begin to cause a decrease in audit quality. As the

upper echelon (e.g. partners) become long-tenured, they are more likely to resist change

and lose fit with the organization (Miller, 1991). When this occurs, the performance of

long-tenured individuals could begin to deteriorate. Many accounting partnerships

structure a mandatory retirement component into their partnership agreements to remove

this performance deterioration effect (Stanger & Carlson, 2017), although the Big 4 do

not have a publically-available official policy regarding this practice. Using Australian

18

partner data, Carey and Simnett (2006) provide some evidence of longer partner tenure

harming audit quality. Audit quality deterioration occurring due to longer tenure may be

more magnified at smaller offices with fewer partners having more individual control,

rather than at larger offices with many partners where control is disseminated (Miller,

1991), meaning if a small office is providing poor audit performance, that performance

would likely exist among a larger portion of the office audit engagements. It might also

be that smaller offices have a single big client or a small handful of big clients, and by

focusing their efforts on those more “important” (i.e., biggest fee) clients, other smaller

audit engagements would suffer. Therefore, this study also proposes that:

H2b: There is a negative relationship between longer audit partner tenure and

audit quality for smaller offices.

Audit Workload Compression and Audit Quality

As previously mentioned, the PCAOB identified auditor workload compression as

an indicator of audit quality (PCAOB, 2015). Workload compression occurs because of

the rush to complete annual report filings subsequent to fiscal year end (i.e. busy season).

The PCAOB considered both partner workload, and manager and staff in their

discussions. A greater client workload might prevent the partner from giving proper

consideration to each audit. Similarly, managers and staff might be pressured to sacrifice

effectiveness for efficiency’s sake, and therefore, not properly plan and perform the

appropriate audit procedures.

When auditors have a higher workload volume, they will be required to make

more judgments and decisions. Behavioral decision theory suggests that auditors make

decisions that harmonize with their core expectations and standards (Slovic, Fischoff, &

19

Lichtenstein, 1977). When auditors gain experience and knowledge, their decisions

become more practical and appropriate (Einhorn & Hogarth, 1981). As auditors become

required to make faster decisions, however, they often perceive fewer environmental cues

that can help to clarify issues and create appropriate responses (Bruner, Matter, &

Papanek, 1955; Easterbrook, 1959; Edwards, 1961). Time compressions due to the busy

filing season require auditors to make faster decisions and can affect audit performance

and audit precision (McDaniel, 1990).

Audit quality issues related to workload compression have been shown to occur at

smaller offices for three primary reasons. First, although higher overall audit workloads

typically occur at offices with more fees, more SEC clients, and more total clients

(DeAngelo, 1981; Reynolds & Francis, 2000; Choi et al., 2010; Francis et al., 2013;

Francis & Yu, 2009), smaller offices have fewer resources and often require auditors to

work on multiple client engagements (Lopez & Peters, 2012). When auditors work on

multiple engagements and multiple engagement types simultaneously, they may

incorrectly analogize audit information from one client to the next (Brown, Gissel, &

Neely, 2016). Second, large clients and SEC clients tend to be more complex both in

their operational activities and their business structures, and often incur significant

litigation risk (DeAngelo, 1981; Simunic, 1980; Hay et al., 2006). The large offices that

support complex clients have greater on-site access to expert auditors than small offices.

When small offices are faced with complex situations, they may be required to consult

with experts at other offices, leading to magnified engagement timing issues. Third,

when SEC clients are regarded as high-risk for PCAOB inspections, offices might

allocate more experienced auditors to those engagements (Moroney, Knechel, &

Dowling, 2017). This allocation might disproportionately affect small offices by causing

20

their more experienced auditors to shoulder a heavier workload. Therefore, this study

hypothesizes that:

H3a: There is a negative relationship between audit workload compression and

audit quality.

It is probable that the available human capital of the office relative to audit

workload compression will also affect audit quality. Audit partners at offices with less

audit human capital may participate more in the management of the audit engagement,

whereas audit partners at offices with more audit human capital may be involved in

highlevel review of the audit engagement (Lopez & Peters, 2012). Similarly, managers

and staff at offices with less audit human capital might be required to work on audit

engagements in multiple industries (e.g. transportation and manufacturing), while

managers and staff from offices with more audit human capital might have expertise in

particular areas (Francis & Yu, 2009). Overall, offices with fewer audit human capital

may be spread too thin such that the greater workload may influence the quality of the

audit. Therefore, this study hypothesizes that:

H3b: The negative relationship between audit workload compression and audit

quality is greater at offices that have relatively less audit human capital.

CHAPTER 3

METHODOLOGY

The Client Sample

To test the relationship between Big 4 office personnel and audit quality, two

separate datasets are obtained – one containing U.S. client data and one containing U.S.

Big 4 employee current and historical information. The sample of audit client data used

to measure audit quality (restatements, internal control weaknesses, going concern

modifications, qualified opinions, and audit fees) covers the period 2011 to 2018, based

on Compustat year descriptions. This range in evaluation years for the client data is to

mitigate the potential gamesmanship that could occur, for example, if the audit firm were

to discount fees for new clients or load negative findings into loss years and final contract

years (Ruiz-Barbadillo, Gomez-Aguilar, & Carrera, 2009). Using this range also

captures effects from post-Dodd-Frank implementation, as many auditors from the

LinkedIn database would have been working since before 2010 when the Dodd-Frank

legislation became effective. The client data is from audit clients of all U.S. Big 4 firms

and was obtained by matching and merging Audit Analytics and Compustat data.

Consistent with prior studies on audit quality, financial institutions and utilities are

excluded from the client sample (e.g. Francis & Yu, 2009; Hay et al., 2006; Francis et al.,

2013). The initial client sample is 79,286 firm-year observations for fiscal years ending

during 2011 through 2018. After excluding duplicates and unmatched clients between

22

Compustat and Audit Analytics (30,737); clients audited by non-Big 4, non-U.S. Big 4,

and non-U.S. clients (27,560); clients audited by U.S. Big 4 offices with unavailable

office personnel information (293); clients from financial institutions and utility

industries (8,175); and remaining clients with missing data (1,220); the final sample

comprises 11,301 observations (See Panel A of Table 2). Panel B of Table 2 breaks down

the client sample by year, and Panel 3 of Table 2 shows the industry membership of the

sample based on two-digit SIC codes.

The Office Personnel Sample

The office personnel composition for the Big 4 firms is captured by accessing

public LinkedIn profiles and extracting and organizing the data into a database. The

public LinkedIn profiles contain variables such as current office, professional level,

experience, education, and gender. As a completeness test, the number of LinkedIn

profiles for the Big 4 firms is compared to the firms’ PCAOB annual reports (see Table 3,

Panel A). Registered public accounting firms that audit publicly traded companies are

required by the PCAOB to file a report that accumulates certain annual activity. Annual

activity includes audits issued, the physical location of firm offices, and the number of

total employees. The evaluation in Panels B and C of Table 3 reveals that a generally

comparable amount of total profiles exist for the Big 4 firms in total (61.75%) and

separately (Big 4 Firm A – 43.10%; Big 4 Firm B – 73.09%; Big 4 Firm C – 79.08%;

Big 4 Firm D – 74.11%).

23

TABLE 2

Client Sample Selection and Industry Membership

Panel A: Client Sample Selection

Observations

Clients listed in Compustat with fiscal years ending

from 2011 to 2018 79,286

Less clients unmatched between Audit Analytics

and Compustat (30,737)

Less clients audited by non-Big 4, non-US Big 4, non-US clients, and

clients that were not audited by

Big 4 after fiscal year ending during 2016 (27,560) Less clients audited by

Big 4 offices with unavailable (293)

information on LinkedIn

Less financial institution and utility clients (8,175)

Less clients with missing data (1,220)

Final Sample 11,301

Panel B: Client Sample by Fiscal Year End

Observations

2011 1,425

2012 1,505

2013

1,609

2014 1,674

2015 1,736

2016 1,787

2017 1,540

2018

25

Panel C: Industry Membership of Client Sample

24

Final Sample

11,301 28 Chemicals and allied products 1,781

15.76

73 Business services 1,466 12.97

36 Electronic and other

electric equipment 817 7.23

38 Instruments and related products 678 6.00

35 Industrial machinery and equipment 658 5.82

13 Oil and gas extraction 521 4.61

37 Transportation equipment 341 3.02

48 Telecommunications 304 2.69

50 Wholesale trade - durable goods 292 2.58

20 Food and kindred products 282 2.50

59 Miscellaneous retail 247 2.19

58 Eating and drinking places 244 2.16

51 Wholesale trade - nondurable goods 221 1.96

56 Apparel and accessory stores 211 1.87

80 Health services 208 1.84

87 Engineering, accounting, research,

management, and related services 200 1.77

21 specific industries 8,471 74.97 Others 44 other industries

2,830 25.03

Total Sample 65 industries 11,301 100.00

Two-Digit

Number

% of

SIC Codes Industry Name

of Firms

Sample

25

TABLE 3

Personnel Sample Selection

Panel A: Personnel Sample Selection

Observations

Total profiles on LinkedIn extracted from

November 2017 to January 2018 206,721

Less profiles with missing data, duplicate profiles, and obvious fake

profiles (66,855) Final Personnel Sample

139,866

Total U.S. Big 4 firm headcount

per PCAOB 2018 annual reports 226,516

Final Personnel Sample as a percent of

PCAOB 2018 annual reports 61.75

Panel B: Composition of U.S. Big 4 Personnel Sample

Observations

Big 4 Firm A 41,406

Big 4 Firm B 37,171

Big 4 Firm C 36,568

Big 4 Firm D 24,721

Final Personnel Sample 139,866

26

Panel C: Comparison of Big 4 Personnel Sample to PCAOB

2018 Annual Firm Reports

Observations

as a % of PCAOB PCAOB

Report Report

Observations Headcount Headcount

Panel A of Table 4 shows the total firm staffing percentages based on the

personnel sample. It should be noted that the totals for each firm include non-audit

personnel, and certain service lines, such as information technology consulting or

valuations, may not have the same staffing proportions as a typical financial statement

audit service line. Panel B of Table 4 shows just the auditor staffing percentages based

on the personnel sample. These percentages align more closely to the descending slope

shape that one would expect to see among audit departments, with audit partners, audit

principals, and audit senior managers constituting the smaller headcount at the top which

Big 4 Firm A

41,406

96,063

43.10

Big 4 Firm B

37,171

50,856

73.09

Big 4 Firm C

36,568

46,240

79.08

Big 4 Firm D

24,721

33,357

74.11

139,866

226,516

61.75

27

then significantly increases when adding layers of audit managers, audit senior

associates, and audit associates.

TABLE 4

Personnel Sample Staffing Proportions

Panel A: U.S. Big 4 Total Staffing % Based on Personnel Sample

Big 4 Big 4

Firm A Firm A

Headcount %

Partners 1,484 3.58

Principals 3,154 7.62

Senior Managers 3,753 9.06

Managers 9,096 21.97

Senior Associates 11,395 27.52

Associates 11,056 26.70

Admin/Legal 1,468 3.55

41,406 100.00

Big 4 Big 4

-

2,000

4,000

6,000

8,000

10,000

12,000

Partners

Principals

Senior Managers

Managers

Senior Associates

Associates

Admin/Legal

Total Firm Headcount

-

Big 4 Firm A

Firm B Firm B

Headcount %

Partners

1,818

4.89

Principals 5,527 14.87

Senior Managers 1,045 2.81

Managers 7,725 20.78

Senior Associates 8,215 22.10

Associates 11,821 31.80

Admin/Legal 1,020 2.74

37,171 100.00

Partners

1,360

5.50

Principals

3,781

15.29

Senior Managers

1,207

4.88

Managers

5,134

20.77

Senior Associates

5,444

22.02

Associates

7,051

28.52

Admin/Legal

744

3.01

24,721

100.00

Big 4

Firm C

Big 4

Firm C

Headcount

%

-

2,000

4,000

6,000

8,000

10,000

12,000

Partners

Principals

Senior Managers

Managers

Senior Associates

Associates

Admin/Legal

Total Firm Headcount

-

Big 4 Firm B

Total Firm Headcount - Big 4 Firm C

Partners

2,005

5.48

Principals

3,511

9.60

Senior Managers

3,256

8.90

Managers

7,639

20.89

Senior Associates

9,463

25.88

Associates

9,621

26.31

Admin/Legal

1,073

2.93

36,568

100.00

Big 4

Firm D

Big 4

Firm D

Headcount

%

-

2,000

4,000

6,000

8,000

10,000

Partners

Principals

Senior Managers

Managers

Senior Associates

Associates

Admin/Legal

Total Firm Headcount - Big 4 Firm D

Partners

6,667

4.77

Principals

15,973

11.42

Senior Managers

9,261

6.62

Managers

29,594

21.16

Senior Associates

34,517

24.68

Associates

39,549

28.28

Admin/Legal

4,305

3.08

139,866

100.00

Total Firm Headcount - All U.S.

Big 4

Total

Headcount

%

-

1,000

2,000

3,000

4,000

5,000

6,000

7,000

8,000

Partners

Principals

Senior Managers

Managers

Senior Associates

Associates

Admin/Legal

-

5,000

10,000

15,000

20,000

25,000

30,000

35,000

40,000

Partners

Principals

Senior Managers

Managers

Senior Associates

Associates

Admin/Legal

Panel B: U.S. Big 4 Auditor Staffing %

Based on Personnel Sample

Total Auditor Headcount

- Big 4 Firm A

- 500 1,000 1,500 2,000 2,500 3,000

Big 4

Firm A

Big 4

Firm A

Headcount

%

Audit Partners

242

3.94

Audit Principals

166

2.70

Audit Senior Managers

475

7.73

Audit Managers

1,074

17.48

Audit Senior Associates

2,811

45.75

Audit Associates

1,376

22.40

6,144

100.00

Big 4

Firm B

Big 4

Firm B

Headcount

%

Audit Partners

452

5.26

Audit Principals

274

3.19

Audit Senior Managers

285

3.32

Audit Managers

1,288

15.00

Audit Senior Associates

1,952

22.73

Audit Associates

4,336

50.49

8,587

100.00

Audit Partners

Audit Principals

Audit Senior Managers

Audit Managers

Audit Senior Associates

Audit Associates

Total Auditor Headcount - Big 4 Firm B

- 500 1,000 1,500 2,000 2,500 3,000 3,500 4,000 4,500

Audit Partners

328

5.10

Audit Principals

251

3.90

Audit Senior Managers

408

6.34

Audit Managers

1,074

16.70

Audit Senior Associates

1,893

29.43

Audit Associates

2,478

38.53

6,432

100.00

Big 4

Firm C

Big 4

Firm C

Headcount

%

Audit Partners

Audit Principals

Audit Senior Managers

Audit Managers

Audit Senior Associates

Audit Associates

Total Auditor Headcount - Big 4

Firm C

- 500 1,000 1,500 2,000 2,500 3,000 3,500

Total Auditor Headcount - Big 4 Firm D

- 500 1,000 1,500 2,000 2,500

Audit Partners

358

4.16

Audit Principals

212

2.46

Audit Senior Managers

524

6.09

Audit Managers

1,447

16.82

Audit Senior Associates

2,667

31.01

Audit Associates

3,393

39.45

8,601

100.00

Big 4

Firm D

Big 4

Firm D

Headcount

%

Audit Partners

Audit Principals

Audit Senior Managers

Audit Managers

Audit Senior Associates

Audit Associates

Audit Partners

Audit Principals

Audit Senior Managers

Audit Managers

Audit Senior Associates

Audit Associates

Audit Partners

1,380

4.64

Audit Principals

903

3.03

Audit Senior Managers

1,692

5.68

Audit Managers

4,883

16.41

Audit Senior Associates

9,323

31.32

Audit Associates

11,583

38.92

29,764

100.00

Total Auditor Headcount - All U.S. Big 4

- 2,000 4,000 6,000 8,000 10,000 12,000

Total

Headcount

%

Audit Partners

Audit Principals

Audit Senior Managers

Audit Managers

Audit Senior Associates

Audit Associates

37

To test the validity of the LinkedIn personnel data sample, certain characteristics

of the Big 4 offices are compared to various data sources. First, individual data provided

by a U.S. Big 4 office is compared to the personnel sample (See Panel A of Table 5). The

total headcount provided by a single U.S. Big 4 office is 1,909 compared to the

LinkedIn personnel headcount for that same office of 1,045. Panel B of Table 5 shows

631 specific names can be matched between the U.S. Big 4 office and LinkedIn, a

60.38% match. A less than full match of names is expected because the U.S. Big 4 office

provides the official first and last name, whereas the personnel sample from LinkedIn

include profiles where the individual uses a maiden/married last name, a “nickname” as a

first name, or a middle name as a first name. Panel C of Table 5 shows that, of the 631

names validated from the U.S. Big 4 data to the LinkedIn personnel sample, 87.80% of

job titles are matched.

TABLE 5

Comparison of Personnel Sample to U.S. Big 4 Office Data

Panel A: Comparison of a List of Employees with Job Titles

Validated by a U.S. Big 4 Office to the Personnel Sample

Headcount Total

Headcount per Headcount per Big 4 % of Personnel % of % Office Total

Sample Total Validated

Partners

168

8.80

45

4.31

26.79

Principals

219

11.47

128

12.25

58.45

Senior Managers

249

13.04

80

7.66

32.13

Managers

282

14.77

163

15.60

57.80

Senior Associates

507

26.56

272

26.03

53.65

Associates

396

20.74

338

32.34

85.35

Admin/Legal

88

4.61

19

1.82

21.59

38

Total

1,909

100.00

1,045

100.00

54.74

Panel B: Names from the Personnel Sample

Matched to the U.S. Big 4 Office List

Headcount Names

Total per Matched

Names

Personnel to Big 4 %

Sample List Validated

1,045 631 60.38

Panel C: Validated Names from the Personnel Sample

Matched to Job Titles from the U.S. Big 4 Office List

Names Job

Matched Matched Titles to Big 4 Job % List Titles Validated

631 554 87.80

Second, comparisons are also made between the personnel sample and the Form

AP data from the PCAOB website (See Table 6). The full Form AP data contains 2,228

unique partners, with firm and office information, who are responsible for audit

engagements filed between January 31, 2017 and April 30, 2018. The personnel sample

contains 1,380 audit partners, of which 623 are validated to the Form AP name, firm, and

office location, a match of 45.14%. Again, a less than full match of names is expected

because the Form AP data provides the engagement partners’ signed name, whereas the

39

personnel sample from LinkedIn will include profiles where the individual uses a

maiden/married last name, a “nickname” as a first name, or a middle name as a first name.

Additionally, it is assumed that the personnel sample includes audit partners that are not

required to sign the audit report, audit partners that solely provide technical expertise and

guidance, and audit partners that specialize in private companies, and therefore will not

appear in the Form AP data.

TABLE 6

Comparison of Personnel Sample Audit Partners

To PCAOB Form AP Data

Headcount Headcount Total per Matched to Headcount Personnel

Form AP %

Sample Data Validated

Audit Partners 1,380 623 45.14

Third, the comparison of total employee headcount from a Big 4 website to the

headcount from the personnel sample is made (See Table 7A). The proportion of

employees per office location as listed by the Big 4 firm from their 2014 campus

information website is compared to the proportion of employees per office location as

listed in the personnel sample. The graph in Table 7A indicates that a reasonable spread of

total headcount in the personnel sample compared to the Big 4 firm data. All three

comparisons tests indicate a reasonable amount of validity in the personnel sample, and

40

the audit client data and personnel data are merged into one database, where individual

firm offices and auditors are matched with the audit clients they served.

43

38

While the audit client data covers 2011 to 2018, based on Compustat year

descriptions, the LinkedIn personnel data was extracted during a short period of time

(November 2017 to January 2018). An initial concern regarding these time frames is that

LinkedIn personnel data cannot be backfilled to dates prior to 2018, and it is possible that

personnel composition and characteristics will have changed drastically over time. To

address issues of using 2018 LinkedIn personnel data with 2011-2018 client data, two

trend comparisons are made to ascertain the extent to which the audit office personnel

composition changes over time. First, a paper by Hoopes, Merkley, Pacelli, and

Schroeder (2018) uses H-1B visa auditor information from the U.S. Department of Labor

to gather certain compensation data for Big 4 associates, senior associates, and managers

from 2004 to 2013 (See Table 7B). An H-1B visa allows foreign individuals to be

employed in certain types of specialty occupations.

TABLE 7B

H-1B Visa Associate, Senior Associate, and Manager Auditor Observations from 2004-

2013 (Hoopes et al., 2018, p. 46)

Observation Year No. of Observations % of Total Sample

2004

897

7%

2005

1,018

8%

2006

1,239

10%

2007

1,714

13%

2008

1,492

12%

2009

1,292

10%

2010

1,034

8%

2011

1,379

11%

2012

1,442

11%

2013

1,289

10%

Total 12,796 100% As Table 7C shows, the observations of audit

managers, audit senior associates, and audit associates compared to the total size of the

sample remains reasonably consistent both in total number of observations and total

percent of the sample. This persistency is expected due to the fact that there are no

significant changes in the number of audited public companies from year to year (i.e.

historically 9,000 – 10,000 public companies are listed in Compustat for each year).

Therefore, the existence of a significant change in the composition of total auditor

headcount from year to year would be unexpected.

Second, information about new U.S. partners and principals from 2015 to 2017

was obtained from promotion press releases issued on the websites for two of the Big 4

firms. The press release information includes, among several items, the offices in which

the newly minted partners and principals are practicing. Table 8 shows that the combined

number of new partners and principals for the two Big 4 firms by office is reasonably

consistent from 2015 to 2017, indicating that the amounts of partner and principal

promotions per office may not vary significantly over time.

TABLE 7C

New Partners and Principals at Two Big 4 Firms

47

Empirical Models and Variables

Consistent with prior research on audit quality, OLS and logarithmic regression

models are estimated, examining the effect of Big 4 auditor office personnel

characteristics on audit quality (Hay et al., 2006; Hossain et al., 2017; Cahan & Sun,

2015). Table 8 defines main and alternative dependent variables (Panel A), test variables

(Panels B1, B2, and B3), and control variables (Panels C-F). Table 8 also presents the

expected association between all the independent variables and the dependent variables.

Audit Human Capital and Audit Quality

Audit human capital is measured at the office audit department level and for each

type of auditor (partners, principals, managers, and associates). These measurements

indicate the resources available to each office, whether physical resources or peer

resources. Prior research assumes that larger offices with more human capital tend to

have higher audit quality (Fun et al., 2012; Francis et al., 2013; Francis & Yu, 2009; Choi

et al., 2010). AU_DEPT is a measure of the number of total audit department staff

divided by total employees for each auditor office. Because a larger audit department

will have more resources, experience, and expertise, AU_DEPT is expected to positively

impact audit quality, and is the test variable for H1a.

Audit Quality = B0 + B1 AU_DEPT + Control Variables +

Year and Industry Fixed Effects + e (1a)

TABLE 8

Variable Definitions

48

Panel A: Main Dependent and Alternative Dependent Variables

Variable Name Variable Measurement (Source)

RESTATE = 1 if the audit client subsequently restates its current year financial

statements, and 0 otherwise (Audit Analytics);

ICW = 1 if the audit client receives an adverse internal control opinion, and 0

otherwise (Compustat-AUOPIC);

RESTATE_NO_ICW-# = 1 if there is a financial restatement without an ICW, and 0 for a

restatement that is accompanied by a related ICW (Audit Analytics,

Compustat-AUOPIC)

GCM = 1 if the audit client receives a going concern modified opinion, and 0

otherwise (Audit Analytics)

AU_FEES = natural logarithm of the audit fees paid by a client to the auditor (Audit

Analytics)

RESTATE_CE-# = 1 if the financial restatement affected core earnings, and 0 otherwise

(Compustat-AUOPIC, Compustat-NOPI, Compustat-NOPIO);

SEC_REG_FAR-# = 1 if the financial restatement was occurred at an auditor office located in

the same city as one of the twelve SEC regional offices, and 0 otherwise

(web search);

#: These variables are alternative measures and will be employed in supplementary analysis.

Panel B1: Test Variables Measuring Audit Human Capital

Expected

Variable Name Sign Variable Measurement (Source)

AU_DEPT +/- total audit department staff divided by total employees for each

auditor office (web search);

LG_OFFICE_SEN_LVL +/- 1 if the total audit partners, principals, senior managers, and

managers for each auditor office is greater than or equal to the

overall median audit partners, principals, senior managers, and

managers for the audit firm for the year, and 0 otherwise (web

search);

PG_AU_PART +/- the proportion of audit partners at the office with a post-graduate

degree (web search);

UG_MSA_AU_PART +/- the proportion of audit partners at the office that attended an

institution located in the same MSA as the office in which they

work (web search);

UG_ELITE_AU_PART +/- the proportion of audit partners at the office that attended an ivy-

league school, or a top-10 accounting school per US News (web

search);

UG_AACSB_ACC_AU_PART +/- the proportion of audit partners at the office that attended a school

with an accounting department separately accredited by the

AACSB (web search);

49

PG_AU_PRIN +/- the proportion of audit principals at the office with a post-

graduate degree (web search);

UG_MSA_AU_PRIN +/- the proportion of audit principals at the office that attended an

institution located in the same MSA as the office in which they

work (web search);

UG_ELITE_AU_PRIN +/- the proportion of audit principals at the office that attended an

ivy-league school, or a top-10 accounting school per US News

(web search);

UG_AACSB_ACC_AU_PRIN +/- the proportion of audit principals at the office that attended a

school with an accounting department separately accredited by

the AACSB (web search);

PG_AU_SMGR_MGR +/- the proportion of audit senior

managers and managers at the office with a post-graduate degree

(web search);

UG_MSA_AU_SMGR_MGR +/- the proportion of audit senior managers and managers at the

office that attended an institution located in the same MSA as the

office in which they work (web search);

UG_ELITE_AU_SMGR_MGR +/- the proportion of audit senior managers and managers at the

office that attended an ivy-league school, or a top-10

accounting school per US News (web search);

UG_AACSB_ACC_AU_SMGR_MGR +/- the proportion of audit senior managers and managers at the

office that attended a school with an accounting department

separately accredited by the AACSB (web search);

PG_AU_SEN_ASSOC +/- the proportion of audit senior

associates and associates at the office with a post-graduate

degree (web search);

UG_MSA_AU_SEN_ASSOC +/- the proportion of audit senior associates and associates at the

office that attended an institution located in the same MSA as the

office in which they work (web search);

UG_ELITE_AU_SEN_ASSOC +/- the proportion of audit senior associates and associates at the

office that attended an ivy-league school, or a top-10

accounting school per US News (web search);

UG_AACSB_ACC_AU_SEN_ASSOC +/- the proportion of audit senior associates and associates at the

office that attended a school with an accounting department

separately accredited by the AACSB (web search);

Panel B2: Test Variables Measuring Audit Personnel Experience

Expected

Variable Name Sign Variable Measurement (Source)

PART_EXP_DEPT_AVG +/- mean audit partner years of service for each auditor office (web

search);

PRIN_EXP_DEPT_AVG +/- mean audit principal years of service for each auditor office (web

search);

SMGR_MGR_EXP_DEPT_AVG +/- mean audit department senior manager and manager years of service

for each auditor office (web search);

SEN_ASSOC_EXP_DEPT_AVG +/- mean audit department senior associate and associate years of service

for each auditor office (web search);

PART_EXP_FIRM_AVG +/- mean audit partner years of service at the firm for each auditor office

(web search);

PRIN_EXP_FIRM_AVG +/- mean audit principal years of service years of service at the firm for

each auditor office (web search);

50

SMGR_MGR_EXP_FIRM_AVG +/- mean audit department senior manager and manager years of service

at the firm for each auditor office (web search);

SEN_ASSOC_EXP_FIRM_AVG +/- mean audit department senior associate and associate years of service

at the firm for each auditor office (web search);

PART_EXP_TOT_AVG +/- mean audit partner years of service as an auditor for each auditor

office (web search);

PRIN_EXP_TOT_AVG +/- mean audit principal years of service as an auditor for each auditor

office (web search);

SMGR_MGR_EXP_TOT_AVG +/- mean audit department senior manager and manager years of service

as an auditor for each auditor office (web search);

SEN_ASSOC_EXP_TOT_AVG +/- mean audit department senior associate and associate years of service

as an auditor for each auditor office (web search);

PART_TENURE_LONG +/- 1 if the partner overall experience as an auditor is above the median of

all other audit partners in the firm, and 0 otherwise (web search);

SM_OFFICE_CLIENTS +/- 1 if the total audit clients for each auditor office is less than the overall

median for the audit firm for the year, and 0 otherwise (web search);

PRIN_TENURE_LONG-# +/- 1 if the principal overall experience as an auditor is above the median

of all other audit principals in the firm, and 0 otherwise (web search);

PART_TENURE_LONG x +/- 1 if the partner overall experience as an auditor is above the

median of SM_OFFICE_CLIENTS all other audit partners in the firm multiplied by 1 if the total audit

clients for each auditor office is less than the overall median for the

audit firm for the year, and 0 otherwise (web search);

PRIN_TENURE_LONG x +/- 1 if the principal overall experience as an auditor is above the median

SM_OFFICE_CLIENTS-# of all other audit principals in the firm multiplied by 1 if the total audit

clients for each auditor office is less than the overall median for the

audit firm for the year, and 0 otherwise (web search);

#: These variables are alternative measures and will be employed in supplementary analysis.

Panel B3: Test Variables Measuring Audit Workload Compression

Expected

Variable Name Sign Variable Measurement (Source)

BUSY_FYE1 +/- number of clients with fiscal years ending in December for each auditor

office (Compustat-COMN, Compustat-FYR, Audit Analytics, web search);

BUSY_FYE1_PART +/- number of clients with fiscal years ending in December per

partner for each auditor office (Compustat-COMN, Compustat-FYR,

Audit Analytics, web search);

BUSY_FYE1_PRIN +/- number of clients with fiscal years ending in December per

principal for each auditor office (Compustat-COMN, Compustat-FYR,

Audit Analytics, web search);

BUSY_FYE1_SMGR_MGR +/- number of clients with fiscal years ending in December per senior

manager and manager for each auditor office (Compustat-COMN,

Compustat-FYR, Audit Analytics, web search);

51

BUSY_FYE1_SEN_ASSOC +/- number of clients with fiscal years ending in December per senior

associate and associate for each auditor office (Compustat-COMN,

Compustat-FYR, Audit Analytics, web search);

BUSY_FYE2_PART-# +/- number of clients with fiscal years ending between December and

January per partner for each auditor office (Compustat-COMN,

Compustat-FYR, Audit Analytics, web search);

BUSY_FYE2_PRIN-# +/- number of clients with fiscal years ending between December and

January per principal for each auditor office (Compustat-COMN,

Compustat-FYR, Audit Analytics, web search);

BUSY_FYE2_SMGR_MGR-# +/- number of clients with fiscal years ending between December and

January per senior manager and manager for each auditor office

(Compustat-COMN, Compustat-FYR, Audit Analytics, web search);

BUSY_FYE2_SEN_ASSOC-# +/- number of clients with fiscal years ending between December and

January per senior associate and associate for each auditor office

(Compustat-COMN, Compustat-FYR, Audit Analytics, web search);

BUSY_FYE3_PART-# +/- number of clients with fiscal years ending between November and

December per partner for each auditor office (Compustat-COMN,

Compustat-FYR, Audit Analytics, web search);

BUSY_FYE3_PRIN-# +/- number of clients with fiscal years ending between November and

December per principal for each auditor office (Compustat-COMN,

Compustat-FYR, Audit Analytics, web search);

BUSY_FYE3_SMGR_MGR-# +/- number of clients with fiscal years ending between November and

December per senior manager and manager for each auditor office

(Compustat-COMN, Compustat-FYR, Audit Analytics, web search);

BUSY_FYE3_SEN_ASSOC-# +/- number of clients with fiscal years ending between November and

December per senior associate and associate for each auditor office

(Compustat-COMN, Compustat-FYR, Audit Analytics, web search);

BUSY_FYE4_PART-# +/- number of clients with fiscal years ending between November and

January per partner for each auditor office (Compustat-COMN,

Compustat-FYR, Audit Analytics, web search);

BUSY_FYE4_PRIN-# +/- number of clients with fiscal years ending between November and

January per principal for each auditor office (Compustat-COMN,

Compustat-FYR, Audit Analytics, web search);

BUSY_FYE4_SMGR_MGR-# +/- number of clients with fiscal years ending between November and

January per senior manager and manager for each auditor office

(Compustat-COMN, Compustat-FYR, Audit Analytics, web search);

BUSY_FYE4_SEN_ASSOC-# +/- number of clients with fiscal years ending between November and

January per senior associate and associate for each auditor office

(Compustat-COMN, Compustat-FYR, Audit Analytics, web search);

LG_OFFICE_AUD_HC +/- number of clients with fiscal years ending in December for each auditor

office multiplied by 1 if the total audit department staff for each auditor

office is greater than or equal to the overall median for the audit firm for

the year, and 0 otherwise (Audit Analytics, web search);

#: These variables are alternative measures and will be employed in supplementary analysis.

52

Panel C: Common Control Variables

Expected

Variable Name Sign

Variable Measurement (Source)

AUDIT_CLIENTS +/- number of total audit clients divided by the number of total

auditors for each auditor office (Audit Analytics and web search);

SIZE +/- natural logarithm of a client's total assets (Compustat-AT);

LOSS +/- 1 if the client has a net loss in the current period, and 0 otherwise

(Compustat-NI)

EQUITY_MULTIPLIER + ratio of total assets to total equity (Compustat-DT, Compustat-AT)

QUICK - proportion of quick assets to current liabilities (Compustat-ACT,

Compustat-INVT, Compustat-XPP, Compustat-LCT);

OPSEG +/- number of operating segments reported for the client (Compustat-

Segments);

GEOSEG +/- number of geographical segments reported for the client (Compustat-

Segments);

LnPOPULATION - natural logarithm of a MSA population (2010 census data); FIRM_TENURE + 1 if client

tenure with the audit firm is three years or less, and 0

otherwise (Audit Analytics);

NATIONAL_LEADER - 1 if the audit firm receives the most fees in an industry at the national

level, and 0 otherwise (Audit Analytics);

CITY_LEADER - 1 if the audit firm receives the most fees in an industry at the city

level, and 0 otherwise (Audit Analytics);

NONAUDIT +/- natural logarithm of a client's non-audit service fees (Audit

Analytics);

FEM_AUDIT ? proportion of office audit department females tot total office audit

department (web search);

INITIAL +/- 1 if the client switched to a new auditor within the prior two years,

and 0 otherwise (Audit Analytics);

MERGER + 1 if there was a merger activity in the current period, and 0 otherwise

(Compustat-AQA);

Panel D: Control Variables Specific to Restatements

Expected

Variable Name Sign Variable Measurement (Source)

CHANGE_IN_RECEIVABLES + percentage change in accounts receivable from t-1 to t (Compustat-

53

ARTFS);

CHANGE_IN_INVENTORIES + percentage change in inventories from t-1 to t (Compustat-INVT);

CHANGE_IN_CASH_SALES + percentage change in cash sales from t-1 to t (Compustat-REVT,

Compustat-ARTFS);

Panel E: Control Variables Specific to Going Concern Opinion Modifications

Variable Name Sign Variable Measurement (Source)

REPORTLAG + number of days between the end of the fiscal year and the date that

earnings are announced (Audit Analytics);

BANKRUPTCY - probability that bankruptcy will not occur, using the Altman Z-score

(Audit Analytics);

Panel F: Control Variables Specific to Audit Fees

Expected

Variable Name Sign Variable Measurement (Source)

FOREIGN + percentage of sales from non-U.S. localities (Factset Revere)

DCFO + 1 if the client reported negative operating cash flows (Compustat-

OANCF);

GROWTH + percentage increase in total assets from the prior period (Compustat-

AT);

In addition to AU_DEPT, a measure for offices in which the total audit partners,

principals, senior managers, and managers are greater than or equal to the overall median

total audit partners, principals, senior managers, and managers for the audit firm is

calculated (LG_OFFICE_SEN_LVL): these are the test variables for H1b. Although all

54

staffing levels are individually expected to have a positive relationship with audit quality,

the relationship is expected to be greater at more senior levels than at less senior levels.

Audit Quality = B0 + B1 AU_DEPT + B2 LG_OFFICE_SEN_LVL + Control

Variables + Year and Industry Fixed Effects + e (1b)

Auditors with a post-graduate degree might positively impact audit quality more

than auditors with just an undergraduate degree. Similarly, auditors from schools that are

known to produce accountants with higher CPA exam pass rates are viewed as being

foundationally strong investments to employers. Because of the higher turnover rates in

public accounting, firms must be able to maintain the right amount of human capital.

Finally, offices that can fill their human capital needs from local universities should be

able to maintain higher audit quality than offices that have to travel outside their local

area to fill their human capital needs. Measures for the following will create an

interaction effect between educational effects and each staffing level, and represent the

test variables for H1c: the proportion of audit partners, principals, senior

managers/managers, and senior associates/associates at the office with a post-graduate

degree (PG_AU_PART; PG_AU_PRIN; PG_AU_SMGR_MGR; PG_AU_SEN_ASSOC);

the proportion of audit partners, principals, senior managers/managers, and senior

associates/associates at the office that attended an ivy-league school, or a top-10

accounting school per US News (UG_ELITE_AU_PART; UG_ELITE_AU_PRIN;

UG_ELITE_AU_SMGR_MGR; UG_ELITE_AU_SEN_ASSOC); the proportion of audit

partners, principals, senior managers/managers, and senior associates/associates at the

office that attended an institution located in the same MSA as the office in which they

55

work (UG_MSA_AU_PART; UG_MSA_AU_PRIN; UG_MSA_AU_SMGR_MGR;

UG_MSA_AU_SEN_ASSOC); and the proportion of audit partners, principals, senior

managers/managers, and senior associates/associates at the office that attended a school

with an accounting department separately accredited by the AACSB

(UG_AACSB_ACC_AU_PART; UG_AACSB_ACC_AU_PRIN;

UG_AACSB_ACC_AU_SMGR_MGR; UG_AACSB_ACC_AU_SEN_ASSOC).

Audit Quality = B0 + B1 PG_AU_PART + B2 UG_ELITE_AU_PART +

B3 UG_MSA_AU_PART + B4 UG_AACSB_ACC_AU_PART + Control Variables + Year

and Industry Fixed Effects + e (1c)

Audit Quality = B0 + B1 PG_AU_PRIN + B2 UG_ELITE_AU_PRIN +

B3 UG_MSA_AU_PRIN + B4 P UG_AACSB_ACC_AU_PRIN + Control Variables +

Year and Industry Fixed Effects + e (1c)

Audit Quality = B0 + B1 PG_AU_SMGR_MGR +

B2 UG_ELITE_AU_SMGR_MGR + B3 UG_MSA_AU_SMGR_MGR + B4

UG_AACSB_ACC_AU_SMGR_MGR + Control Variables + Year and Industry Fixed

Effects + e (1c)

Audit Quality = B0 + B1 PG_AU_SEN_ASSOC +

B2 UG_ELITE_AU_SEN_ASSOC + B3 UG_MSA_AU_SEN_ASSOC + B4

UG_AACSB_ACC_AU_SEN_ASSOC + Control Variables + Year and Industry Fixed

Effects + e (1c)

56

Audit Personnel Experience and Audit Quality

Auditor experience is measured at three levels (audit office, audit firm, and audit

profession) for each type of auditor (partner, principal, senior manager/manager, and

senior associate/associate). Both Cahan and Sun (2015) and Gul et al., (2013) show the

years of experience of a signing audit partner may have a positive relationship with audit

quality, while Hossain et al. (2015) show that the experience of senior and staff associates

may have a less significant relationship with audit quality. In addition, prior experiments

have shown that more experienced auditors tend to make the most appropriate decisions

(Kinney & Uecker, 1982). Audit office experience is measured as the mean years of

service at the office audit department for each staffing level and is expected to have a

greater positive impact to audit quality at more senior auditor levels than at less senior

auditor levels (PART_EXP_ DEPT_AVG, PRIN_EXP_ DEPT_AVG,

SMGR_MGR_EXP_ DEPT_AVG, and SEN_ASSOC_EXP_ DEPT_AVG), and constitutes

the test variables for H2a (audit department experience).

Audit Quality = B0 + B1 PART_EXP_ DEPT_AVG + B2 PRIN_EXP_ DEPT_AVG

+ B3 SMGR_MGR_EXP_ DEPT_AVG + B4 SEN_ASSOC_EXP_ DEPT_AVG + Control

Variables + Year and Industry Fixed Effects + e (2a)

Auditor firm experience is measured as the mean years of service at the audit

firm, capturing any effects of auditor experience in more than one office of their current

firm, and is expected to have a greater positive impact to audit quality at more senior

auditor levels than at less senior auditor levels (PART_EXP_ FIRM_AVG, PRIN_EXP_

FIRM_AVG, SMGR_MGR_EXP_ FIRM_AVG, and SEN_ASSOC_EXP_ FIRM_AVG),

57

and constitutes the test variables for H2a (audit firm experience).

Audit Quality = B0 + B1 PART_EXP_ FIRM_AVG + B2 PRIN_EXP_ FIRM_AVG

+ B3 SMGR_MGR_EXP_ FIRM_AVG + B4 SEN_ASSOC_EXP_ FIRM_AVG + Control

Variables + Year and Industry Fixed Effects + e (2a)

Auditor total experience is measured as the mean years of service as an auditor,

capturing any effects of auditor experience at different firms, and is expected to have a

greater positive impact to audit quality at more senior auditor levels than at less senior

auditor levels (PART_EXP_ TOT_AVG, PRIN_EXP_ TOT_AVG,

SMGR_MGR_EXP_ TOT_AVG, and SEN_ASSOC_EXP_ TOT_AVG), and constitutes the

test variables for H2a (total auditor experience).

Audit Quality = B0 + B1 PART_EXP_ TOT_AVG + B2 PRIN_EXP_ TOT_AVG +

B3 SMGR_MGR_EXP_ TOT_AVG + B4 SEN_ASSOC_EXP_ TOT_AVG + Control

Variables + Year and Industry Fixed Effects + e (2a)

There is some evidence to suggest that individuals in leadership positions begin to

lose their productivity after reaching an extended level of tenure due to complacency and

resistance to firm changes (Miller, 1991; Carey & Simnett, 2006). Additionally, the

productivity of these individuals may decline more at smaller offices rather than at larger

offices (Miller, 1991). Therefore, this study measures both long and short audit partner

tenure and its effect on audit quality, while also considering the size of the office audit

department. Typically, accounting studies have calculated audit partner tenure with

58

respect to the length of time associated with particular audit clients (Carey & Simnett,

2006). However, since this study is focused more on the partner’s overall personal

experience, this study uses Miller’s (1991) method of dividing CEO tenure subgroups

into long and short according to the median years of experience. The variable

PART_TENURE_LONG equals 1 if the partner’s overall experience as an auditor is above

the median of all other audit partners at the firm, and 0 otherwise.

SM_OFFICE_CLIENTS is 1 if the total audit clients for each auditor office is less than

the overall median for the audit firm for the year, and 0 otherwise, and is a measure of

office audit department size.

PART_TENURE_LONG x SM_OFFICE_CLIENTS is the interaction of these two

variables created to assess main effects. These are the test variables.

Audit Quality = B0 + B1 PART_TENURE_LONG + B2 SM_OFFICE_CLIENTS +

B3 PART_TENURE_LONG x SM_OFFICE_CLIENTS + Control Variables + Year and

Industry Fixed Effects + e (2b)

In supplemental analyses, this study performs the same measurements for

principals. Principals are sometimes considered to be client service partners or nonequity

partners and might also possess a deteriorating audit quality effect of long tenure.

The variable PRIN_TENURE_LONG equals 1 if the principal’s overall experience as an

auditor is above the median, and 0 otherwise. PRIN_TENURE_LONG x

SM_OFFICE_CLIENTS is the interaction of 1 if the principal overall experience as an

auditor is above the median of all other audit principals in the firm multiplied by 1 if the

59

total audit clients for each auditor office is less than the overall median for the audit firm

for the year, and 0 otherwise

Audit Workload Compression and Audit Quality

Auditor workload compression is measured by the size of the audit department in

terms of audit staff and for each type of auditor in the audit department. The largest

portion of audit client fiscal years end during December, resulting in a two to three month

busy season compression. This compression has been shown to result in decreased audit

quality (Lopez & Peters, 2012). The variable BUSY_FYE1 is the number of clients with

fiscal years ending in December for each auditor office, and is the test variable for H3a.

Audit Quality = B0 + B1 BUSY_FYE1 + Control Variables + Year and Industry

Fixed Effects + e (3a)

It is expected that workload compression affects the audit department associate

and manager levels to a greater extent since they are working at the engagement level

rather than the high-level review (Lopez & Peters, 2012). Therefore, this study also

measures the number of December fiscal year-end audit clients as a proportion of the

office audit department staffing level (BUSY_FYE1_PART, BUSY_FYE1_PRIN,

BUSY_FYE1_SMGR_MGR, and BUSY_FYE1_SEN_ASSOC). Because workload

compression is expected to affect large offices differently than smaller offices, this study

also measures the number of clients with fiscal years ending in December for each

auditor office multiplied by 1 if the total audit department staff for each auditor office is

greater than or equal to the overall median for the audit firm for the year, and 0 otherwise

60

(LG_OFFICE_AUD_HC). Together, these constitute the test variables for H3b. In

supplemental analysis, this study also measures these variables for alternate November

(BUSY_FYE2_PART, BUSY_FYE2_PRIN, BUSY_FYE2_SMGR_MGR, and

BUSY_FYE2_SEN_ASSOC), January (BUSY_FYE3_PART, BUSY_FYE3_PRIN,

BUSY_FYE3_SMGR_MGR, and BUSY_FYE3_SEN_ASSOC) and November to January

(BUSY_FYE4_PART, BUSY_FYE4_PRIN, BUSY_FYE4_SMGR_MGR, and

BUSY_FYE4_SEN_ASSOC) busy season windows to investigate the sensitivity of busy

season to audit planning, issues, and delays (Lopez & Peters, 2012).

Audit Quality = B0 + B1 BUSY_FYE1_PART + B2 BUSY_FYE1_PRIN + B3

BUSY_FYE1_SMGR_MGR + B4 BUSY_FYE1_SEN_ASSOC +

B5 LG_OFFICE_AUD_HC + Control Variables + Year and Industry Fixed Effects

+ e (3b)

Dependent Variables

To make the connection between the experience, workload, and audit human

capital and the quality of audits conducted by the engagement office, it is necessary to

discuss several common measures of audit quality that this study examines. The first

measure of audit quality is financial restatements (RESTATE). Restatements of client

financial information have a negative financial reporting quality connotation and can

cause the market to react, leading to auditor litigation (Knechel et al., 2013; Francis,

2004; Stanley & DeZoort, 2007). A restatement of prior period information commonly

occurs due to the correction of a material error and suggests that the auditor failed to

detect and/or report the material error, hence audit quality is low when the audit client

61

restates its financial statements. RESTATE equals 1 if the audit client subsequently

restates its current year financial statements, and 0 otherwise.

The size of the restatement depends not only on the pressure involved but also the

nature of the misstatement. Prior research has shown that restatements affecting core

earnings (operating revenues and operating expenses) tend to be larger than restatements

affecting non-core earnings (non-operating revenues and non-operating expenses) (Hribar

& Jenkins, 2004). Based on the analysis of Francis et al. (2013), the assumption is that

higher quality auditors will have fewer client financial restatements affecting core

earnings. In supplemental analyses, this study measures audit quality by examining

restatements related to core earnings. RESTATE_CE equals 1 if in the restatement is

related to core earnings, and 0 otherwise.

The second measure of audit quality widely used in the literature (Knechel et al.,

2013) is presence of an internal control weakness (ICW). The variable ICW equals 1 if

the client receives an adverse internal control opinion, and 0 otherwise. ICWs exist when

a missing or improperly designed control is so important that its absence could result in a

material financial misstatement. Auditors are required under SOX to report the existence

of ICWs; therefore, audit quality is high when ICWs are detected.

In supplemental analyses and depending on the sample of firms with a financial

restatement, this study utilizes a measure of audit quality that combines the first two

measures. This study employs the presence of a restatement without a preceding internal

control problem identified by the auditor, that is denoted RESTATE_NO_ICW. Since a

financial restatement occurs due to a breakdown in internal controls, a restatement

without a related adverse SOX 404 material weakness opinion has been interpreted to

suggest that the auditor failed to detect or report the material weakness (Rice, Weber, &

62

Biyu, 2015; Rice & Weber, 2012). Hence, RESTATE_NO_ICW equals 1 if there is a

financial restatement without an ICW, and 0 for a restatement that is accompanied by a

related ICW.

The third measure of audit quality relates to the auditor’s opinion. Throughout

the course of fieldwork, auditors are required to determine if the operations of the client

are viable to continue for the foreseeable future (AICPA, 2017). If the auditor determines

that there are going concern issues, then they must modify the audit opinion. Failure to

issue a going concern modified opinion when warranted is considered an audit failure.

The going concern opinion decision is therefore confined to audit clients that are in

distress where distress is defined as negative income and/or negative operating cash

flows. The variable GCM equals 1 if the client receives a going concern modified

opinion, and 0 otherwise.

The final measure of audit quality is audit fees. Prior studies often use audit fees

as a proxy for audit quality justifying that higher audit fees represent additional audit

work (DeAngelo, 1981; Reynolds & Francis, 2000; Choi et al., 2010; Francis & Yu,

2009; Cahan & Sun, 2015). The variable audit fees, AU_FEES, is measured as the

natural logarithm of the audit fees paid by a client to the auditor.

Common Audit Quality Control Variables

The first group of control variables captures client information common to the

RESTATE, ICW, GCM, and AU_FEES (i.e., audit quality) dependent variables, based on

prior studies. The variable SIZE is measured by taking the natural logarithm of total

client assets. Higher values of the variable SIZE should correlate to lower RESTATE,

lower ICW, lower GCM, and higher AU_FEES (Francis et al., 2013; Hossain et al., 2017;

63

Francis & Yu, 2009). The variable LOSS equals 1 if the client has a net loss in the current

period, and 0 otherwise. Higher values of LOSS should correlate to higher RESTATE,

higher ICW, higher GCM, and higher AU_FEES (Francis et al., 2013; Hennes et al.,

2008). INITIAL takes a value of 1 if the client switched to a new auditor within the prior

two years, and 0 otherwise (Rice & Weber, 2012). Higher values of INITIAL should

correlate to higher RESTATE, lower ICW, lower GCM, and lower AU_FEES. MERGER

equals 1 if there was a merger activity during the current period, and 0 otherwise (Stanley

& DeZoort, 2007). Higher values of MERGER should correlate to higher RESTATE,

higher ICW, lower GCM, and higher AU_FEES.

EQUITY_MULTIPLIER is a variable that calculates the ratio of total assets to total equity

(Francis & Yu, 2009). Higher values of EQUITY_MULTIPLIER should correlate to

higher RESTATE, higher ICW, higher GCM, and higher AU_FEES. The variable QUICK

(quick assets divided by current liabilities) is a measure of liquidity (Carey & Simnett,

2006; Reynolds & Francis, 2000). Higher values of QUICK should correlate to lower

RESTATE, lower ICW, lower GCM, and lower AU_FEES. The total number of operating

segments (OPSEG) and the number of geographical segments (GEOSEG) control for

client complexity due to the additional audit effort required. Higher values of GEOSEG

and OPSEG should correlate to higher RESTATE, higher ICW, lower GCM, and higher

AU_FEES. This study will also measure the natural logarithm of metro area population

(LnPOPULATION) because companies located in more populated cities have more

access to individuals with greater accounting expertise (Francis et al., 2005). Higher

values of LnPOPULATION should correlate to lower RESTATE, lower ICW, lower GCM,

and higher AU_FEES. FIRM_TENURE (1 if client tenure with the audit firm is three

years or less, and 0 otherwise) controls for client tenure with the audit firm (Francis &

64

Yu, 2009; Johnson, Khurana, & Reynolds, 2002). Higher values of FIRM_TENURE

should correlate to higher RESTATE, lower ICW, lower GCM, and higher AU_FEES.

Finally, this study includes industry classification codes (INDUSTRY) to control for the

effects of industry type on audit quality. The type of industry could either increase or

decrease audit quality, and thus, no prediction is made to its relationship with audit

quality.

The second group of control variables captures auditor characteristics common to

the measures of audit quality in this study. Auditors that specialize in a particular area are

more likely to provide higher quality audits (and be more expensive than) auditors

without specialization. Prior research shows that audit firms receiving the largest

industry-specific fees at both the national level (NATIONAL_LEADER) and the city level

(CITY_LEADER) provide higher levels of audit quality. Higher values of

NATIONAL_LEADER and CITY_LEADER (each coded as 1 if leader, and 0 otherwise)

should correlate to lower RESTATE, higher ICW, higher GCM, and higher AU_FEES.

NONAUDIT represents the natural logarithm of fees paid to the audit firm for other

services provided. Higher values of NONAUDIT should correlate to higher RESTATE,

higher ICW, higher GCM, and higher AU_FEES. (Rice & Weber, 2012). FEM_AU is a

variable that calculates the proportion of office audit department females to total office

audit department individuals. A small group of studies have investigated non-U.S. Big 4

firms, and found some evidence that clients pay higher fees to auditor offices with larger

proportions of female partners (Hardies et al., 2015; Ittonen et al., 2013). This study

provides the opportunity to examine gender diversity with respect to the existence of a

female audit fee premium in U.S. Big 4 firms. It is unclear to what extent the

composition of males and females in the audit department may impact audit quality

65

because there is no prior evidence. Therefore, this study makes no prediction as to the

direction of the relationship between FEM_AU and audit quality.

Control Variables Specific to Restatements

In addition to the common audit quality control variables previously discussed,

the restatement regression models control for specific client and auditor characteristics

from prior studies that use restatements to proxy for audit quality. Clients with more

accruals have more opportunity to manage earnings and/or misrepresent financial

information (Francis & Yu, 2009; Francis et al., 2013). Higher values of

CHANGE_IN_RECIEVABLES, CHANGE_IN_INVENTORIES, and

CHANGE_IN_CASH_SALES should correlate to higher RESTATE.

Control Variables Specific to Going Concern Opinion Modifications

The internal control weakness regression models control for specific client and

auditor characteristics from prior studies that use going concern opinion modifications to

proxy for audit quality. A number of variables used in prior studies for going concern

testing indicate clients that are financially distressed. REPORTLAG measures the

number of days between the end of the fiscal year and the date that earnings are

announced. Higher REPORTLAG should correlate to higher GCM (Francis & Yu, 2009;

Reynolds & Francis, 2000). Finally, BANKRUPTCY measures the probability that

bankruptcy will not occur, and is calculated using the Altman Z-score (Francis & Yu,

2009; Altman, 1983). Higher BANKRUPTCY should correlate to lower GCM.

66

Control Variables Specific to Audit Fees

The AU_FEES regression models control for specific client and auditor

characteristics from prior studies that use audit fees to proxy for audit quality. FOREIGN

(measured as the percentage of sales from non-U.S. localities), are often more complex,

and therefore require higher audit fees. Higher FOREIGN should correlate to higher

AU_FEES (Hay et al., 2006; Choi et al., 2010). DCFO equals 1 if the client reported

negative operating cash flows, and 0 otherwise. Higher DCFO should correlate to higher

AU_FEES; GROWTH measures the percentage increase in total assets from the prior

period. Higher GROWTH should correlate to higher AU_FEES.

CHAPTER 4

DATA ANALYSIS AND FINDINGS

Descriptive Statistics

Panel A of Table 9 reports descriptive statistics for the measures of audit quality

across the full sample for fiscal years ending during 2011 through 2018. Companies

within the full sample have a mean (median) RESTATE of 0.091 (0.000), ICW of 0.027

(0.000), GCM of 0.022 (0.000), and AU_FEES of 14.374 (14.319), all of which are

comparable with prior studies (for example, Francis et al., 2013; Francis & Yu, 2009;

Hossain et al., 2017). Panel B1 of Table 9 reports descriptive statistics for the test

variables measuring audit human capital for fiscal years ending during 2011 through

2018 and they first show that, as the mean of LG_OFFICE_SEN_LVL is 0.810, a large

proportion of the client sample is audited by Big 4 offices with high concentrations of

senior-level audit human capital; and second, more managers and associates have

postgraduate degrees than partners and principals, more associates are employed from

schools in the same MSA as their home office, and that all other education-related

variables are reasonably consistent at all staffing levels. Panel B2 of Table 9 reports

descriptive statistics for the test variables measuring audit personnel experience for fiscal

years ending during 2011 through 2018 and they show that, as expected, the average

68

61

years of service at the office (DEPT), firm (FIRM), and as a professional (TOTAL) for

more experienced staffing levels such as PART_EXP_DEPT_AVG,

PART_EXP_FIRM_AVG, and PART_EXP_TOT_AVG have much higher mean (median)

values of 13.184 years, 14.550 years, and 15.299 years (13.390 years, 14.898 years, and

15.765 years), respectively, compared to the average years of service at the office

(DEPT), firm (FIRM), and as a professional (TOTAL) for less experienced staffing levels

such as SEN_ASSOC_EXP_DEPT_AVG, SEN_ASSOC_EXP_FIRM_AVG, and

SEN_ASSOC_EXP_TOT_AVG with mean (median) values of 2.352 years, 2.732 years,

and 2.815 years (2.305 years, 2.653 years, and 2.722 years), respectively. Panel B3 of

Table 9 reports descriptive statistics for the test variables measuring audit workload

compression for fiscal years ending during 2011 through 2018 and they show that, as

expected, senior associates and associates (BUSY_FYE1_SEN_ASSOC) have much more

client availability due to their sheer numbers (0.210 clients per audit senior

associate/associate mean and 0.175 client per audit senior associate/associate median),

than partners (BUSY_FYE1_PART), who typically will be much less in number (3.643

clients per partner mean and 0.175 clients per partner median). Descriptive statistics for

control variables for fiscal years ending during 2011 through 2018 are presented in Panel

C of Table 9 and are also comparable with prior studies (for example, Francis et al.,

2013; Francis & Yu, 2009; Hossain et al., 2017).

Multivariate Analysis - Correlations

Table 10 reports the results of Person correlation matrices for the full 2011-2018

client sample. Panel A of Table 10 reports correlations of test variables measuring audit

69

human capital and control variables (Hypotheses 1a, 1b, and 1c), and while there are a

number of significant correlations between the variables measuring education at the

staffing levels, these will be test in separate regressions, thereby eliminating the

multicollinearity threat. The largest VIF (untabulated) is lower than the common

tolerance value of 10 at which the risk of errors due to multicollinearity increases (Hair et

al., 2010). Panel B of Table 11 reports correlations of test variables measuring audit

personnel experience and control variables (Hypotheses 2a and 2b). The audit personnel

experience test variables are naturally highly correlated; however, they are tested in

separate regressions thereby posing no risk of error due to multicollinearity. There are a

number of other significant correlations, but none are large enough to be considered a

multicollinearity threat. The largest VIF (not including test variables) is lower than 10

(untabulated). Panel C of Table 11 reports correlations of test variables measuring audit

workload compression and control variables (Hypotheses 3a and 3b). There are a number

of other significant correlations, but none are large enough to be considered a

multicollinearity threat. The largest VIF is lower than 10 (untabulated). Panel D of Table

11 reports correlations of control variables. There are a number of other significant

correlations, but none are large enough to be considered a multicollinearity threat. The

largest VIF is lower than 10 (untabulated).

TABLE 9

Panel A: Descriptive Statistics for Main Dependent Variables

Full Sample (n = 11,301)

Mean Median Std. Dev. Q1

Q3

70

RESTATE 0.091 0.000 0.288 0.000

0.000

ICW 0.027 0.000 0.163 0.000

0.000

GCM 0.022 0.000 0.146 0.000

0.000

AU_FEES 14.374 14.319 1.029 13.670

15.027

Panel B1: Descriptive Statistics for Test Variables Measuring Audit Human Capital

Full Sample (n = 11,301)

Mean

Median

Std. Dev.

Q1

Q3

AU_DEPT

0.278

0.279

0.105

0.206

0.313

LG_OFFICE_SEN_LVL

0.810

1.000

0.393

1.000

PG_AU_PART

0.225

0.204

0.169

0.122

0.333

UG_MSA_AU_PART

0.240

0.192

0.100

0.346

UG_ELITE_AU_PART

0.121

0.063

0.173

0.000

0.182

UG_AACSB_ACC_AU_PART

0.181

0.154

0.185

0.000

0.250

PG_AU_PRIN

0.297

0.311

0.236

0.125

0.385

UG_MSA_AU_PRIN

0.280

0.308

0.212

0.111

0.375

UG_ELITE_AU_PRIN

0.097

0.000

0.172

0.000

0.143

UG_AACSB_ACC_AU_PRIN

0.153

0.100

0.218

0.000

0.222

PG_AU_SMGR_MGR

0.389

0.409

0.183

0.317

0.484

UG_MSA_AU_SMGR_MGR

0.276

0.269

0.172

0.192

0.365

UG_ELITE_AU_SMGR_MGR

0.077

0.043

0.101

0.000

0.107

UG_AACSB_ACC_AU_SMGR_MGR

0.163

0.140

0.143

0.060

0.211

PG_AU_SEN_ASSOC

0.519

0.504

0.121

0.433

0.599

UG_MSA_AU_SEN_ASSOC

0.357

0.346

0.140

0.292

0.407

UG_ELITE_AU_SEN_ASSOC

0.079

0.045

0.089

0.023

0.106

UG_AACSB_ACC_AU_SEN_ASSOC

0.168

0.137

0.121

0.079

0.220

Panel B2: Descriptive Statistics for Test Variables Measuring Audit Personnel Experience

Full Sample (n = 11,301)

Mean Median Std. Dev.

PART_EXP_DEPT_AVG 13.184 13.390 5.100 11.509

14.972

PRIN_EXP_DEPT_AVG 7.594 7.380 4.617 5.004

10.007

71

Descriptive Statistics

Panel B3: Descriptive Statistics for Test Variables Measuring Audit Workload Compression

Full Sample (n = 11,301)

Mean Median Std. Dev. Q1 Q3

BUSY_FYE1

BUSY_FYE1_PART

BUSY_FYE1_PRIN

BUSY_FYE1_SMGR_MGR

BUSY_FYE1_SEN_ASSOC

LG_OFFICE_AUD_HC

SMGR+MGR_EXP_DEPT_AVG 6.028 4.962 6.019 4.535

5.804

SEN_ASSOC_EXP_DEPT_AVG 2.352 2.305 0.280 2.182

2.454

PART_EXP_FIRM_AVG 14.550 14.898 5.062 12.816

16.703

PRIN_EXP_FIRM_AVG 9.487 9.170 5.949 6.226

11.508

SMGR+MGR_EXP_FIRM_AVG 7.012 5.906 6.966 5.204

6.893

SEN_ASSOC_EXP_FIRM_AVG 2.732 2.653 0.448 2.390

3.062

PART_EXP_TOT_AVG 15.299 15.765 5.163 13.244

17.277

PRIN_EXP_TOT_AVG 9.862 9.791 6.002 6.782

11.908

SMGR+MGR_EXP_TOT_AVG 7.208 6.149 6.961 5.382

7.079

SEN_ASSOC_EXP_TOT_AVG 2.815 2.722 0.453 2.479

3.115

PART_TENURE_LONG 0.440 0.000 0.496 0.000

1.000

SM_OFFICE_CLIENTS 0.180 0.000 0.385 0.000

0.000

PART_TENURE_LONG x SM_OFFICE_CLIENTS 0.367 0.000 0.482 0.000

1.000

Q1

Q3

72

Panel C: Descriptive Statistics for Control

Variables

Mean

Median

Std. Dev. Q1 Q3

AUDIT_CLIENTS

0.256

0.181

0.255

0.134

0.294

SIZE

20.979

21.008

1.865

19.685

22.216

LOSS

0.310

0.000

0.463

0.000

1.000

LOSS_NCF

0.328

0.000

0.469

0.000

1.000

EQUITY_MULTIPLIER

2.521

1.979

8.243

1.436

2.865

QUICK

2.398

1.443

3.060

0.966

2.472

OPSEG

1.310

0.000

4.094

0.000

GEOSEG

6.860

5.000

7.299

2.000

9.000

lnPOPULATION

14.857

14.876

0.836

14.224

15.392

FIRM_TENURE

0.100

0.000

0.305

0.000

NATIONAL_LEADER

0.390

0.000

0.488

0.000

1.000

CITY_LEADER

0.640

1.000

0.479

0.000

1.000

NONAUDIT

11.117

12.226

4.051

10.597

13.469

FEM_AU

0.468

0.476

0.058

0.435

0.500

INITIAL

0.070

0.000

0.258

0.000

MERGER

0.370

0.000

0.483

0.000

1.000

CHANGE_IN_RECEIVABLES

-0.070

-0.058

11.501

-0.161

0.005

CHANGE_IN_INVENTORIES

0.015

0.000

1.423

-0.077

0.000

CHANGE_IN_REVENUE

0.567

0.055

22.099

-0.005

0.170

REPORTLAG

63.110

59.000

29.106

53.000

67.000

BANKRUPTCY

3.918

3.052

7.384

1.520

4.966

FOREIGN

0.218

0.002

0.773

0.000

0.373

DCFO

0.160

0.000

0.369

0.000

GROWTH

0.198

0.043

0.750

-0.023

0.158

46.970

34.000

45.207

13.000

68.000

3.643

2.538

3.728

1.683

4.182

5.990

4.000

6.091

2.375

7.571

0.933

0.663

1.228

0.419

0.910

0.210

0.175

0.136

0.129

0.255

0.823

1.000

0.381

Full Sample (n = 11,301)

1.000

TABLE 10 - Correlations

Panel A: Correlations of Test Variables Measuring Audit Human Capital and Control Variables

AU_DE

PT

LG_OF

FICE_S

EN_LVL

PG_AU

_PART

UG_MS

A_AU_P

ART

UG_ELI

TE_AU_

PART

UG_AA

CSB_AC

C_AU_

PART

PG_AU

_PRIN

UG_MS

A_AU_P

RIN

UG_ELI

TE_AU_

PRIN

UG_AA

CSB_AC

C_AU_

PRIN

PG_AU

_SMGR

_MGR

UG_MS

A_AU_S

MGR_M

GR

UG_ELI

TE_AU_

SMGR_

MGR

UG_AA

CSB_AC

C_AU_S

MGR_M

GR

PG_AU

_SEN_A

SSOC

UG_MS

A_AU_S

EN_ASS

OC

UG_ELI

TE_AU_

SEN_AS

SOC

UG_AA

CSB_AC

C_AU_S

EN_ASS

OC

AU_DEPT

1.000

LG_OFFICE_SEN_LVL

-.434**

1.000

PG_AU_PART

-.111**

.091**

1.000

UG_MSA_AU_PART

.019*

.102**

-.126**

1.000

UG_ELITE_AU_PART

-.040**

-.097**

.056**

-.004

1.000

UG_AACSB_ACC_AU_PART

-.023*

-.112**

.123**

-.120**

.224**

1.000

PG_AU_PRIN

-.154**

-.075**

.196**

-.117**

.020*

.038**

1.000

UG_MSA_AU_PRIN

.026**

.113**

-.186**

.330**

-.079**

-.211**

.042**

1.000

UG_ELITE_AU_PRIN

-.085**

-.064**

.108**

-.129**

.249**

.088**

.114**

.013

1.000

UG_AACSB_ACC_AU_PRIN

.050**

-.287**

.053**

.007

.301**

.254**

.122**

.010

.292**

1.000

PG_AU_SMGR_MGR

.083**

-.198**

.119**

-.146**

.095**

.000

-.019

.008

.158**

.020*

1.000

UG_MSA_AU_SMGR_MGR

.131**

-.098**

-.125**

.206**

-.114**

-.227**

-.134**

.183**

.028**

-.039**

.351**

1.000

UG_ELITE_AU_SMGR_MGR

-.173**

-.008

.128**

.018

.426**

.066**

-.077**

.061**

.402**

.116**

.144**

.039**

1.000

UG_AACSB_ACC_AU_SMGR_MGR

-.175**

-.231**

.090**

-.175**

.214**

.314**

.073**

.017

.099**

.211**

.309**

.099**

.371**

1.000

PG_AU_SEN_ASSOC

.063**

-.148**

.179**

.090**

.011

.004

.182**

-.010

.119**

.081**

.455**

.163**

.147**

.223**

1.000

UG_MSA_AU_SEN_ASSOC

.185**

-.174**

-.213**

.419**

-.106**

-.250**

-.048**

.166**

-.086**

-.066**

.109**

.621**

.013

-.040**

.264**

1.000

UG_ELITE_AU_SEN_ASSOC

-.200**

.100**

.069**

.197**

.560**

.002

.053**

.200**

.456**

.198**

.006

-.018

.633**

.180**

.085**

-.026**

1.000

UG_AACSB_ACC_AU_SEN_ASSOC

-.108**

-.242**

.147**

-.116**

.249**

.324**

.190**

.007

.124**

.331**

.264**

-.021*

.328**

.701**

.370**

-.005

.284**

1.000

AUDIT_CLIENTS

-.022*

.035**

-.079**

.091**

.015

-.187**

-.118**

.158**

.109**

-.177**

.098**

.114**

.109**

-.095**

-.057**

-.001

.164**

-.140**

SIZE

-.070**

-.023*

.100**

-.129**

.044**

.095**

.086**

-.061**

.078**

.026**

.064**

-.057**

.069**

.086**

.056**

-.088**

.031**

.090**

LOSS

.016

.086**

-.072**

.104**

-.066**

-.058**

-.034**

.060**

-.060**

-.064**

-.114**

-.010

-.056**

-.077**

-.026**

.046**

-.005

-.055**

EQUITY_MULTIPLIER

-.016

-.020*

.006

-.015

.003

-.002

-.008

.000

.009

-.015

-.011

-.004

.004

.010

-.010

-.012

.003

.000

QUICK

.023*

.097**

-.056**

.135**

-.055**

-.074**

.045**

-.077**

-.057**

-.110**

.005

-.075**

-.118**

-.039**

.044**

-.027**

-.111**

OPSEG

.010

-.037**

.037**

-.085**

.026**

-.015

.040**

-.024*

.035**

.008

.035**

-.004

.042**

.039**

.025**

-.027**

.022*

-.003

GEOSEG

-.042**

.015

-.007

-.025*

.023*

-.049**

-.008

-.001

.030**

-.025*

-.022*

-.009

.029**

-.018

.000

-.015

.008

-.002

lnPOPULATION

-.277**

.250**

.108**

.023*

.069**

-.084**

.027**

-.029**

.146**

-.081**

.176**

.204**

.142**

.055**

.115**

.034**

.207**

-.046**

FIRM_TENURE

-.031**

.025**

-.003

.038**

-.012

-.003

.019

.015

-.002

-.023*

-.027**

-.009

-.005

-.016

.008

.017

-.006

-.016

NATIONAL_LEADER

.084**

-.019*

.018

.105**

.028**

-.014

-.039**

-.020*

-.065**

.047**

.005

.000

-.040**

-.016

.067**

.028**

.069**

.009

CITY_LEADER

.134**

-.110**

-.057**

-.008

.017

.072**

-.099**

-.001

-.051**

-.035**

.111**

.087**

-.003

.065**

.039**

.053**

-.004

.045**

NONAUDIT

-.081**

.015

.053**

-.098**

.020*

.043**

-.007

-.045**

.030**

-.018

.034**

-.038**

.012

.001

-.050**

-.065**

-.018

-.017

FEM_AU

-.174**

.161**

-.030**

.048**

-.025**

-.202**

-.098**

-.089**

-.080**

-.263**

-.015

-.021*

-.131**

-.168**

.037**

.018

-.009

-.210**

INITIAL

-.030**

.029**

.003

.031**

-.016

-.009

.016

.012

-.016

-.019

-.039**

-.015

-.008

-.013

.002

.015

-.006

-.007

MERGER

-.052**

.012

.018

-.016

.020*

-.025**

.015

-.007

.030**

-.008

.003

-.003

.028**

-.002

-.015

.007

.009

-.006

CHANGE_IN_RECEIVABLES

-.002

.002

-.005

-.007

.008

.004

.010

-.005

.008

.003

-.004

-.001

.003

.007

.003

.007

.009

CHANGE_IN_INVENTORIES

.004

.003

.020*

-.011

-.001

.015

-.010

-.002

-.004

-.001

-.016

-.006

-.009

-.010

-.011

-.006

-.014

CHANGE_IN_REVENUE

-.005

.004

-.003

.000

.016

.000

.002

.000

.001

-.008

-.004

-.011

.003

.000

.003

-.009

.004

-.008

REPORTLAG

-.009

.018

-.030**

.027**

-.038**

-.032**

-.030**

.002

-.044**

-.038**

.007

-.023*

-.036**

-.021*

.017

-.020*

-.031**

BANKRUPTCY

.016

.008

.010

.043**

.004

-.016

-.025*

.010

-.033**

-.018

-.044**

-.002

-.014

-.023*

-.028**

.016

.002

-.027**

FIOREIGN

-.028**

.020*

-.001

-.024*

.000

-.036**

.009

.001

.008

-.006

-.022*

-.015

-.002

-.023*

-.027**

.004

-.002

-.025**

DCFO

.024*

.084**

-.095**

.128**

-.079**

-.075**

-.044**

.071**

-.078**

-.045**

-.114**

-.002

-.084**

-.106**

-.045**

.050**

-.032**

-.087**

GROWTH

.004

.042**

-.015

.047**

-.027**

-.024*

.019

-.026**

-.039**

-.045**

-.004

-.028**

-.053**

-.012

.021*

-.011

-.046**

Panel B: Correlations of Test Variables Measuring Audit Personnel Experience and Control Variables

PRIN_

EXP_T

OT_AV

G

SMGR

_MGR

_EXP_

TOT_A

VG

SEN_A

SSOC_

EXP_T

OT_AV

G

PART_

TENU

RE_LO

NG

SM_O

FFICE

_CLIE

NTS

PART_

TENU

RE_LO

NG

xSM_O

FFICE

_CLIE

NTS

PART_

EXP_D

EPT_A

VG

PRIN_

EXP_D

EPT_A

VG

SMGR

_MGR

_EXP_

DEPT_

AVG

SEN_A

SSOC_

EXP_D

EPT_A

VG

PART_

EXP_F

IRM_A

VG

PRIN_

EXP_F

IRM_A

VG

SMGR_

MGR_E

XP_FIR

M_AVG

SEN_A

SSOC_

EXP_F

IRM_A

VG

PART_

EXP_T

OT_AV

G

PART_EXP_DEPT_AVG

1.000

PRIN_EXP_DEPT_AVG

.157**

1.000

SMGR_MGR_EXP_DEPT_AVG

.017

.106**

1.000

SEN_ASSOC_EXP_DEPT_AVG

.108**

.114**

.083**

1.000

PART_EXP_FIRM_AVG

.945**

.160**

-.003

.111**

1.000

PRIN_EXP_FIRM_AVG

.166**

.700**

.137**

.036**

.198**

1.000

SMGR+MGR_EXP_FIRM_AVG

.026**

.109**

.995**

.088**

.009

.153**

1.000

SEN+ASSOC_EXP_FIRM_AVG

.056**

.066**

.015

.622**

.123**

.179**

.035**

1.000

PART_EXP_TOT_AVG

.924**

.122**

-.023*

.064**

.971**

.148**

-.012

.080**

1.000

PRIN_EXP_TOT_AVG

.171**

.708**

.158**

.021*

.199**

.993**

.174**

.169**

.152**

1.000

SMGR_MGR_EXP_TOT_AVG

.025**

.108**

.994**

.086**

.009

.152**

1.000**

.036**

-.011

.174**

1.000

SEN_ASSOC_EXP_TOT_AVG

.062**

.065**

.016

.601**

.129**

.167**

.037**

.991**

.088**

.159**

.040**

1.000

PART_TENURE_LONG

.579**

-.004

-.102**

.028**

.611**

.111**

-.087**

.047**

.656**

.106**

-.086**

.047**

1.000

SM_OFFICE_CLIENTS

PART_TENURE_LONG x

-.172**

-.092**

-.051**

.130**

-.177**

-.138**

-.055**

-.086**

-.187**

-.150**

-.059**

-.128**

-.034**

1.000

SM_OFFICE_CLIENTS

.325**

.072**

-.021*

.051**

.304**

-.004

-.024*

-.082**

.321**

-.002

-.026**

-.103**

.316**

.598**

1.000

AUDIT_CLIENTS

.069**

.034**

-.075**

.082**

.089**

.011

-.068**

.028**

.112**

.008

-.073**

.012

.048**

-.139**

-.103**

SIZE

-.083**

-.047**

-.065**

-.017

-.079**

-.019*

-.057**

.060**

-.082**

-.029**

-.057**

.055**

-.089**

.036**

-.017

LOSS

.037**

.030**

.052**

-.032**

.043**

.019*

.048**

-.016

.047**

.027**

.049**

-.007

.062**

-.078**

-.040**

EQUITY_MULTIPLIER

-.014

-.008

-.001

.028**

-.012

-.001

.000

.021*

-.016

-.003

.000

.017

-.012

.026**

.011

QUICK

.043**

.041**

.053**

-.066**

.052**

.048**

.049**

-.080**

.058**

.049**

-.071**

.073**

-.089**

-.043**

OPSEG

-.020*

-.053**

-.033**

.002

-.009

-.026**

-.032**

.011

-.022*

-.031**

-.032**

.007

-.058**

.040**

-.016

GEOSEG

.001

.019*

-.003

-.024*

.001

.044**

.002

.004

-.003

.042**

.001

.008

-.002

-.019*

-.032**

lnPOPULATION

.119**

.080**

-.161**

.039**

.116**

-.013

-.147**

.127**

.112**

-.006

-.146**

.151**

.065**

-.166**

-.065**

FIRM_TENURE

.021*

.018

.038**

.002

.024*

.012

.038**

.005

.023*

.014

.038**

.006

.022*

-.021*

-.017

NATIONAL_LEADER

-.023*

.018

-.054**

-.133**

-.014

-.021*

-.061**

-.125**

-.009

-.022*

-.058**

-.117**

.012

-.004

-.003

CITY_LEADER

-.006

-.022*

-.113**

-.024*

-.019*

-.010

-.113**

-.123**

-.018

-.016

-.113**

-.128**

-.011

.090**

.050**

NONAUDIT

-.042**

.007

-.011

-.001

-.030**

.041**

.000

.027**

-.042**

.039**

-.001

.032**

-.061**

.003

-.011

FEM_AU

.090**

.043**

.114**

.178**

.100**

-.062**

.116**

.166**

.125**

-.065**

.114**

.177**

.101**

-.116**

.031**

INITIAL

.016

.004

.042**

-.002

.018

-.003

.041**

.001

.018

-.001

.041**

.005

.020*

-.032**

-.025**

MERGER

-.023*

.002

.018

-.008

-.019*

.014

.021*

.004

-.027**

.014

.021*

.009

-.040**

-.015

-.023*

CHANGE_IN_RECEIVABLES

-.006

.001

.000

.001

-.004

.004

-.001

.017

-.005

.003

-.001

.015

-.012

-.002

-.003

CHANGE_IN_INVENTORIES

-.006

-.002

.001

.002

-.003

.011

.001

-.004

-.003

.011

-.002

-.009

CHANGE_IN_REVENUE

.013

.002

-.003

-.005

.007

.002

-.003

-.009

.010

.005

-.002

-.008

.011

-.008

-.006

REPORTLAG

.014

.005

.035**

.012

-.006

.034**

-.008

.013

-.002

.034**

-.002

.021*

-.025**

-.009

BANKRUPTCY

.029**

.025**

.051**

-.021*

.035**

.048**

.050**

-.034**

.034**

.049**

.050**

-.033**

.023*

-.024*

-.013

FIOREIGN

.007

.004

.024*

.002

.006

.018

.026**

.012

.000

.021*

.025**

.013

.001

-.014

.002

DCFO

.039**

.038**

.051**

-.031**

.042**

.027**

.046**

-.049**

.046**

.037**

.046**

-.040**

.071**

-.076**

-.032**

GROWTH

.020*

.008

.023*

-.024*

.021*

.011

.023*

-.026**

.020*

.014

.023*

-.022*

.027**

-.040**

-.021*

Panel C: Correlations of Test Variables Measuring Audit Workload Compression and Control Variables

BUSY_FYE1_ BUSY_FYE1_ LG_OFFICE

BUSY_FYE1

BUSY_FYE1_PART

BUSY_FYE1_PRIN

SMGR_MGR

SEN_ASSOC

_AUD_HC

BUSY_FYE1

1.000

BUSY_FYE1_PART

.486**

1.000

BUSY_FYE1_PRIN

.208**

.272**

1.000

BUSY_FYE1_SMGR_MGR

-.046**

.062**

.208**

1.000

BUSY_FYE1_SEN_ASSOC

.026**

.337**

.417**

.244**

1.000

LG_OFFICE_AUD_HC

.400**

-.029**

.182**

-.068**

-.177**

1.000

SIZE

-.015

-.009

-.097**

-.039**

-.054**

-.019*

LOSS

.078**

-.027**

.096**

.013

.015

.079**

EQUITY_MULTIPLIER

-.002

.001

-.014

.014

.003

-.014

QUICK

.123**

.033**

.129**

.031**

.032**

.080**

OPSEG

-.013

.029**

-.019*

-.032**

.011

-.028**

GEOSEG

.029**

.014

-.003

.011

.005

.009

lnPOPULATION

.154**

.030**

.013

-.145**

-.020*

.259**

FIRM_TENURE

.003

-.021*

.014

.012

-.003

.012

NATIONAL_LEADER

.017

.016

.097**

-.038**

-.011

-.020*

CITY_LEADER

-.134**

-.015

.034**

-.052**

.050**

-.098**

NONAUDIT

.069**

.039**

-.001

-.006

-.043**

.036**

FEM_AU

.176**

.023*

-.030**

.067**

.049**

.096**

INITIAL

.022*

-.022*

.013

.019*

.003

.018*

MERGER

.011

-.013

-.024*

.033**

-.027**

.020*

CHANGE_IN_RECEIVABLES

.009

.007

-.002

.000

.001

CHANGE_IN_INVENTORIES

.014

.004

.002

-.007

-.005

.003

CHANGE_IN_REVENUE

-.005

.000

-.003

-.009

.011

REPORTLAG

.057**

.007

.037**

.031**

.018

.020*

BANKRUPTCY

.001

.005

.030**

.029**

-.007

.005

FOREIGN

.023*

.031**

-.003

.015

-.003

.016

DCFO

.099**

-.002

.117**

.018

.029**

.071**

GROWTH

.051**

-.002

.051**

.018

.007

.037**

Panel D: Correlations of Control Variables

AUDIT_

CLIENT

S

SIZE

LOSS

EQUITY

_MULTI

PLIER

QUICK

OPSEG

GEOSE

G

lnPOPU

LATION

FIRM_T

ENURE

NATION

AL_LEA

DER

CITY_L

EADER

NONAU

DIT

FEM_A

U

INITIAL

MERGE

R

CHANG

E_IN_R

CHANG

E_IN_I

NVENT

CHANG

E_IN_R

EVENU

REPOR

TLAG

ECEIVA

BLES

ORIES

E

AUDIT_CLIENTS

1.000

SIZE

LOSS

-.048** .013

1.000

-.410**

1.000

EQUITY_MULTIPLIER

.006

.100**

-.033**

1.000

QUICK

.041**

-.354**

.277**

-.059**

1.000

OPSEG

GEOSEG

.020*

.025**

.218**

.211**

-.096**

-.161**

.034**

.003

-.110**

-.112**

1.000

.102**

1.000

lnPOPULATION

.103**

.030**

-.012

.001

-.015

.047**

.017

1.000

FIRM_TENURE

.028**

-.139**

.143**

-.007

.098**

-.047**

-.092**

.007

1.000

NATIONAL_LEADER

CITY_LEADER

-.058** -.012

-.063**

.097**

.070**

-.058**

.004

.019*

.065**

-.066**

-.014

.081**

-.067**

-.039**

.047**

-.051**

.006

-.041**

1.000

.228**

1.000

NONAUDIT

.006

.479**

-.210**

.059**

-.207**

.137**

.214**

.048**

-.097**

-.054**

.058**

1.000

FEM_AU

.198**

-.020*

.054**

.001

.045**

-.020*

.001

.061**

.011

-.022*

-.085**

-.020*

1.000

INITIAL

MERGER

.024* .004

-.174**

.199**

.162**

-.120**

-.024**

.031**

.105**

-.126**

-.048**

.021*

-.085**

.111**

-.004

.033**

.490** -

.012

.013

-.044**

-.035** -

.014

-.132**

.179**

.014

-.009

1.000

-.040**

1.000

CHANGE_IN_RECEIVABLES

-.004

.004

-.002

-.006

.003

.000

.004

-.001

.009

.012

.006

.001

.005

.013

1.000

CHANGE_IN_INVENTORIES

-.008

.005

.004

.009

-.004

.000

-.008

-.002

.014

-.001

-.020*

.009

.001

-.007

.004

-.003

1.000

CHANGE_IN_REVENUE

-.006

-.028**

.010

-.001

.016

-.007

-.020*

-.004

.009

.001

-.002

-.006

-.001

.014

.015

-.001

1.000

REPORTLAG

.031**

-.358**

.216**

-.040**

.066**

-.060**

-.099**

-.005

.141**

.032**

-.029**

-.200**

.021*

.308**

-.077**

.006

.002

.022*

1.000

BANKRUPTCY

.009

.008

-.142**

-.018

.469**

-.029**

.005

-.035**

.023*

-.012

-.019*

-.025**

-.017

.008

-.016

-.011

-.005

-.006

-.153**

FIOREIGN

-.003

.128**

-.133**

.004

-.035**

.049**

.192**

.012

-.024*

-.032**

-.026**

.117**

-.009

-.026**

.079**

-.001

-.007

.000

-.052**

DCFO

.034**

-.465**

.555**

-.047**

.405**

-.108**

-.176**

-.023*

.132**

.097**

-.050**

-.232**

.038**

.130**

-.178**

.014

-.008

.028**

.207**

GROWTH

.012

-.140**

.137**

-.015

.325**

-.052**

-.097**

-.012

.171**

.034**

-.021*

-.107**

.008

.272**

.031**

-.009

-.001

.052**

.116**

Panel D: Correlations of Control Variables (continued)

BANKRUPTCY FIOREIGN DCFO GROWTH

BANKRUPTCY 1.000

FIOREIGN .006 1.000

DCFO -.057** -.098** 1.000

GROWTH .198** -.024** .176** 1.000

*, **, *** denote significance at 0.10, 0.05, and 0.01 levels, respectively. See Table 8 for variable definitions.

70

Logistic and OLS Regressions – Audit Human Capital (H1a, H1b, H1c)

Audit Human Capital at the Audit Office Department and Audit Quality (H1a)

Tables 11 through 19 present the regression results of the relationships between

the test variables and audit quality. Each table includes control variables based on 2011-

2018 client data, as well as control variables based on 2017-2018 client data for purposes

of sensitivity analysis, since the personnel data was extracted from November 2017 to

January 2018. Table 11 provides the regression results of the relationship between audit

human capital and audit quality for Hypothesis 1a. Panel A of Table 11 presents the

results from the logistic regression of RESTATE on the test and control variables. The

coefficients for year and industry fixed effects are not reported for brevity in this table

and all regression tables. The results do not indicate a significant association between

RESTATE and AU_DEPT (p = 0.312). This result suggests that, alternative to H1a, audit

quality measured by client restatements is not impacted by Big 4 offices with greater

proportions of audit human capital. However, the variable capturing the number of total

audit clients at each office as a proportion of total auditors at each office,

(AUDIT_CLIENTS), has a significant negative association with RESTATE (p = 0.020).

This result suggests that, in general, offices with more clients per auditor have fewer

restatements, possibly due to increased exposure from client-related risks, or because

these offices are better able to scale audit resources, such as audit workpaper templates,

planning documents, and internal control narratives for audit clients that operate in

similar industries. The results also show expected significant positive relationships

between RESTATE and ICW, AU_FEES, EQUITY_MULTIPLIER, and FIRM_TENURE,

71

and significant negative relationships with SIZE, NATIONAL_LEADER and QUICK.

Because the personnel data sample also contains gender information, an opportunity is

provided to ascertain the impact of gender on audit quality. Interestingly, the results

indicate a significant negative association between RESTATE and FEM_AU (p = 0.009),

suggesting that audit departments with greater amounts of female auditors have fewer

client restatements, and therefore higher audit quality. When evaluating the results using

only 2017-2018 control variable years, the test variable is not significant, and none of the

control variable results are unexpected.

Panel B of Table 11 presents the results from the logistic regression of ICW on the

test and control variables for Hypothesis 1a. The coefficients for year and industry fixed

effects are not reported. The results indicate a significant positive relationship between

ICW and AU_DEPT (p = 0.088), suggesting that the quantity of audit human capital

available to the office positively impacts audit quality measured by the detection of

material weaknesses. The significant positive relationship between ICW and AU_FEES

is expected because, consistent with the literature, as auditors uncover material

weaknesses, they expand testing and incur additional fees. In addition, the significant

positive relationships between ICW and MERGER and LOSS are expected because,

consistent with the literature, complexities and losses may result in risks that are not

adequately addressed by internal control activities.

TABLE 11 - Regression of Audit Quality on Audit Human Capital at the Audit Office Department (H1a)

Panel A: Using RESTATE as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Variable Wald

Intercept

?

-2.219

3.370

0.066*

-22.338

0.000

0.999

AU_DEPT

-

-0.179

0.242

0.312

-0.570

0.240

0.312

AUDIT_CLIENTS

-

-0.327

4.224

0.020**

-0.131

0.097

0.378

ICW

+

1.409

107.612

0.000***

2.144

30.776

0.000***

AU_FEES

+

0.329

23.215

0.000***

0.276

1.637

0.101

SIZE

-

-0.166

17.101

0.000***

-0.202

2.537

0.056*

LOSS

+

0.004

0.002

0.484

0.486

3.129

0.077*

EQUITY_MULTIPLIER

+

0.007

3.829

0.025**

0.004

0.122

0.364

QUICK

-

-0.023

2.167

0.071*

-0.049

0.754

0.193

OPSEG

+

0.006

0.454

0.250

0.029

0.232

0.315

GEOSEG

+

0.001

0.049

0.412

0.007

0.029

0.433

LnPOPULATION

-

0.000

0.497

0.157

1.491

0.222

FIRM_TENURE

+

0.234

3.784

0.026**

0.133

0.128

0.360

NATIONAL_LEADER

-

-0.148

4.138

0.021**

-0.094

0.182

0.335

Expected Sign

Estimate

Wald

p

-

value

Estimate

p

-

value

CITY_LEADER

-

0.014

0.035

0.851

0.200

0.752

0.386

NONAUDIT

+

0.009

0.722

0.198

0.031

0.856

0.178

FEM_AU

?

-1.617

6.859

0.009***

-2.211

1.307

0.253

INITIAL

+

-0.007

0.002

0.962

0.712

2.132

0.072*

MERGER

+

-0.012

0.029

0.866

0.090

0.171

0.340

CHANGE_IN_RECEIVABLES

+

-0.001

0.122

0.727

-0.019

0.017

0.898

CHANGE_IN_INVENTORIES

+

-0.004

0.025

0.874

-0.005

0.010

0.919

CHANGE_IN_CASH_SALES

+

-0.002

1.019

0.313

-0.071

1.757

0.185

Year and Industry fixed effects

Yes

Number of Restatements

1,025

122

Total Number of Observations

11,301

1,524

Model p-value

0.000***

Pseudo R2

5.6%

19.0%

Regression of Audit Quality on Audit Human Capital at the Audit Office Department (H1a)

Panel B: Using ICW as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Expected Sign

Estimate

Wald

p

-

value

Estimate

p

-

value

Variable Wald

Intercept ? -26.612 0.000 0.998 -21.295 0.000 0.999

AU_DEPT + 0.802 1.838 0.088* 0.596 0.132 0.359

AUDIT_CLIENTS + 0.145 0.519 0.236 -0.945 1.032 0.310

AU_FEES + 1.445 142.247 0.000*** 1.241 11.259 0.000***

SIZE - -0.797 124.839 0.000*** -0.791 13.561 0.000***

LOSS + 0.193 1.829 0.088* 0.079 0.043 0.418

EQUITY_MULTIPLIER + 0.006 0.916 0.170 -0.012 0.557 0.455

QUICK - -0.046 1.859 0.087* -0.214 1.997 0.079*

OPSEG + -0.010 0.372 0.542 0.002 0.000 0.493

GEOSEG + 0.001 0.007 0.468 -0.029 0.163 0.687

LnPOPULATION - 0.066 0.822 0.364 -0.137 0.462 0.249

FIRM_TENURE - 0.245 1.481 0.224 0.612 1.399 0.237

NATIONAL_LEADER + -0.110 0.708 0.400 0.099 0.082 0.388

CITY_LEADER + -0.148 1.163 0.281 0.179 0.242 0.312

NONAUDIT + 0.004 0.042 0.419 -0.005 0.011 0.917

FEM_AU ? 1.562 2.219 0.136 3.191 1.016 0.313

INITIAL - -0.331 1.698 0.193 0.128 0.035 0.851

MERGER + 0.265 4.316 0.019** 0.516 2.250 0.067*

Year and Industry fixed effects Yes Yes

Number of Control Weaknesses 308 45

Total Number of Observations 11,301 1,565

Model p-value 0.000*** 0.113

Pseudo R2 13.3% 23.2%

Regression of Audit Quality on Audit Human Capital at the Audit Office Department (H1a)

Panel C: Using GCM as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Variable Wald p-value Wald

Intercept

?

10.374

13.536

0.000***

10.426

0.000

1.000

AU_DEPT

+

-0.921

0.914

0.339

-4.398

1.256

0.262

AUDIT_CLIENTS

-

-0.223

0.358

0.275

1.141

0.932

0.334

ICW

+

0.140

0.093

0.380

-15.539

0.000

0.998

AU_FEES

+

0.168

1.518

0.109

0.103

0.025

0.437

SIZE

-

-0.458

24.098

0.000***

-0.779

2.950

0.043**

LOSS

+

0.367

0.495

0.241

17.875

0.000

0.499

EQUITY_MULTIPLIER

+

-0.003

0.130

0.718

-0.026

0.166

0.683

QUICK

-

-0.162

18.381

0.000***

-0.096

0.918

0.169

OPSEG

-

0.012

0.095

0.758

-0.184

0.300

0.292

Expected Sign

Estimate

p

-

value

GEOSEG

-

-0.051

5.320

0.021**

0.191

2.161

0.142

LnPOPULATION

-

-0.179

2.507

0.057*

-0.885

4.379

0.018**

FIRM_TENURE

-

0.332

2.059

0.151

0.668

0.679

0.410

NATIONAL_LEADER

+

-0.292

2.371

0.124

-1.070

2.509

0.113

CITY_LEADER

+

0.162

0.730

0.197

0.592

0.744

0.195

NONAUDIT

+

0.001

0.004

0.474

-0.064

1.309

0.253

FEM_AU

?

2.155

2.087

0.149

-5.955

1.228

0.268

INITIAL

-

-0.239

0.858

0.177

-0.843

0.585

0.444

MERGER

-

-0.451

2.907

0.044**

-0.222

0.060

0.806

REPORTLAG

+

0.002

1.366

0.122

0.054

4.003

0.023**

BANKRUPTCY

-

-0.069

34.707

0.000***

-0.076

3.243

0.036**

Year and Industry fixed effects

Yes

Number of Going Concern Modifications

235

19

Total Number of Observations

3,705

481

Model p-value

0.000***

Pseudo R2

40.0%

58.8%

Regression of Audit Quality on Audit Human Capital at the Audit Office Department (H1a)

Panel D: Using AU_FEES as the Dependent Variable

Control Variable Years: 2011-2018

Control Variable Years: 2017-2018

Variable

t-value

p-value

t-value

p-value

Intercept

?

3.981

31.183

0.000***

4.513

13.816

0.000***

AU_DEPT

+

-0.403

-8.216

0.000***

-0.435

-3.481

0.001***

AUDIT_CLIENTS

+

-0.165

-8.630

0.000***

-0.150

-3.222

0.001***

ICW

+

0.343

11.836

0.000***

0.241

3.338

0.000***

RESTATE

+

0.084

5.100

0.000***

0.071

1.574

0.058*

GCM

+

0.117

3.486

0.000***

0.233

2.721

0.004***

SIZE

+

0.462

128.382

0.000***

0.447

47.609

0.000***

LOSS

+

0.168

13.023

0.000***

0.130

4.006

0.000***

EQUITY_MULTIPLIER

+

0.000

0.848

0.199

0.001

0.976

0.165

QUICK

-

-0.025

-13.278

0.000***

-0.025

-4.575

0.000***

OPSEG

+

0.012

9.421

0.000***

0.032

4.330

0.000***

GEOSEG

+

0.008

10.987

0.000***

0.027

5.458

0.000***

lnPOPULATION

+

0.025

4.141

0.000***

0.022

1.426

0.077*

FIRM_TENURE

+

0.057

3.263

0.000***

-0.026

-0.603

0.546

NATIONAL_LEADER

+

-0.011

-1.101

0.271

0.009

0.357

0.361

CITY_LEADER

+

0.034

3.207

0.000***

0.039

1.457

0.073*

NONAUDIT

+

0.017

12.494

0.000***

0.016

4.837

0.000***

FEM_AU

?

0.517

6.058

0.000***

0.307

1.410

0.159

INITIAL

-

0.100

4.684

0.000***

0.078

1.223

0.222

FOREIGN

+

0.050

7.967

0.000***

0.018

1.384

0.084*

MERGER

+

0.053

5.086

0.000***

0.059

2.276

0.012**

DCFO

+

0.115

6.634

0.000***

0.125

2.714

0.004***

GROWTH

+

0.014

1.974

0.024**

0.019

0.879

0.190

Year and Industry fixed effects

Yes

Total Number of Observations

11,301

1,083

Model p-value

0.000***

R2

77.1%

78.8%

Expected Sign

Estimate

90

The significant negative relationships between ICW and SIZE and QUICK are also

expected because, consistent with the literature, larger and/or liquid clients have greater

flexibility to develop and maintain high quality systems of internal controls. When

evaluating the results using only 2017-2018 control variable years, the model becomes

insignificant, possibly due to the small incidence of ICWs.

Panel C of Table 11 presents the results from the logistic regression of GCM on

the test and control variables for Hypothesis 1a. The results do not indicate a significant

relationship between GCM and AU_DEPT (p = 0.339), suggesting that the quantity of

audit human capital available to the office may not impact audit quality measured by

going concern modifications of client financial statements. The highly significant

negative relationship between GCM and BANKRUPTCY is expected, as BANKRUPTCY

measures the Altman-Z Score, where higher numbers indicate greater financial health and

lower likelihood of going concern issues. The significant negative relationships between

GCM and SIZE, QUICK, GEOSEG, LnPOPULATION, and MERGER are also expected

because, consistent with the literature, larger clients, more liquid clients, and clients with

greater market spreads, and engaging in a merger are less likely to exhibit indicators of

failure. When evaluating the results using only 2017-2018 control variable years, the test

variable is not significant, and none of the control variable results are unexpected.

Panel D of Table 11 presents the results from the OLS regression of AU_FEES on

the test and control variables for Hypothesis 1a. The results indicate a significant

negative relationship between AU_FEES and AU_DEPT (p < 0.01), opposite of the

91

positive relationship that was hypothesized in H1a. This finding might indicate that

offices with more available human capital are more productive and efficient. This

finding is strengthened by the significant negative relationship between AU_FEES and

AUDIT_CLIENTS (p < 0.01), which is also opposite of expectation, yet indicates that

offices with more clients per auditor are able to generate “economies of scale” through

resource sharing to control fees. As expected, based on prior literature, the findings

indicate significant positive relationships between AU_FEES and ICW, RESTATE, SIZE,

LOSS, OP_SEG, GEO_SEG, LnPOPULATION, FIRM_TENURE, CITY_LEADER (when

controlling for NATIONAL_LEADER), NONAUDIT, FOREIGN, MERGER, DCFO, and

GROWTH, and a significant negative relationship between AU_FEES and QUICK. In

addition, the results indicate the presence of a significant female audit fee premium as

there is a positive relationship between AU_FEES and FEM_AU (p < 0.01), perhaps

suggesting that offices with greater proportions of female auditors command a fee

premium due to greater audit quality. The results unexpectedly note a significant positive

relationship between AU_FEES and INITIAL, as a “low-ball” effect commonly occurs in

the first year of an audit. However, as INITIAL measures whether a client switched

auditors in the prior two years, it might be that the new auditor is “catching up” on fees in

the years immediately subsequent to the switch. When evaluating the results using only

2017-2018 control variable years, the findings also indicate a significant negative

relationship between AU_FEES and AU_DEPT (p < 0.01), while none of the control

variable results are unexpected.

92

Results Summary for Audit Human Capital at the Audit Office Department and Audit

Quality (H1a)

The test results for H1a overall are mixed, which makes it difficult to draw any

solid conclusions on the effect of human capital on audit quality. The results first make

no indication that audit offices with greater amounts of human capital have fewer

subsequent client financial restatements or greater going concern opinion modifications,

which do not align with H1a. In addition, the results show a negative relationship

between audit department human capital and audit fees, which also does not align with

H1a, yet do show a significant relationship between the amount of audit department

human capital and the reporting of material internal control weaknesses which does align

with H1a. These results could indicate that offices with the greatest amount of available

audit human capital (or biggest audit offices) possess more high quality auditors willing

to report material internal control weaknesses, and are able to manage audit fees more

effectively than offices with lesser amounts of available audit human capital (smaller

audit offices). When evaluating the test results for H1a using only control variables from

2017-2018, the extraction dates for the personnel sample, the test variable results align

with the full sample results (2011-2018), and none of the control variable results are

unexpected.

Supplemental Analysis for Audit Human Capital at the Audit Office Department and

93

Audit Quality (H1a)

In supplemental analysis of H1a testing on the full sample, the dependent variable

RESTATE is modified in three ways. First, RESTATE is replaced with

RESTATE_NO_ICW, to test for restatements in which the auditor failed to report a

material internal control weakness. The model for this test (untabulated) is not

significant (p = 0.603), likely because there are so few instances of restatements

unaccompanied by a material internal control weakness (n = 41). Second, RESTATE is

replaced by RESTATE_CE, to test for restatements only affecting core earnings. The

model for this test (untabulated) is significant (p = 0.012), with a pseudo R-square of

0.132; however, similar to the main model in Panel A of Table 11, the test variable

AU_DEPT remains insignificant (p = 0.143). Finally, RESTATE is replaced by

SEC_REG_FAR, to test for restatements that occurred at an auditor office in the same city

as one of the eleven SEC regional offices. The model for this test (untabulated) is

significant (p < 0.01), with a pseudo R-square of 0.088, and produces a significant

negative relationship between the test variable (AU_DEPT: p < 0.01) and the dependent

variable (SEC_REG_FAR), indicating that audit departments with more available human

capital might be less likely to issue restated financial statements when they are located at

offices in the same city as one of the eleven SEC regional offices, possibly due to the

perceived exposure to regulatory oversight experienced by the auditor and/or client.

Senior-Level Audit Human Capital at the Audit Office Department and Audit Quality

(H1b)

94

Table 12 provides the regression results of the relationship between audit human

capital and audit quality for Hypothesis 1b. Panel A of Table 12 presents the results from

the logistic regression of RESTATE on the test and control variables. The coefficients for

industry and year fixed effects for all regressions are not tabulated to conserve space.

The results indicate a significant negative association between RESTATE and AU_DEPT

(p = 0.087), a significant negative relationship between RESTATE and

LG_OFFICE_SEN_LVL (p = 0.006). These results appear to align with H1b in that

offices with more senior-level audit human capital, rather than just more overall audit

human capital, seem to have fewer future restatements. The results also indicate

significant positive relationships between RESTATE and ICW, AU_FEES,

EQUITY_MULTIPLIER, and FIRM_TENURE, and significant negative relationships

between RESTATE and AUDIT_CLIENTS, NATIONAL_LEADER, QUICK, and SIZE, all

of which are expected based on the prior literature. Finally, the results indicate a

significant negative association between RESTATE and FEM_AU (p = 0.016), suggesting

that audit departments with greater amounts of female auditors have fewer future client

financial restatements, adding to the results from Panel A of Table 11. When evaluating

the results using only 2017-2018 control variable years, the findings indicate a significant

negative relationship between RESTATE and LG_OFFICE_SEN_LVL (p = 0.083), while

none of the control variable results are unexpected.

Panel B of Table 12 presents the results from the logistic regression of ICW on the

test and control variables for Hypothesis 1b. The results do not indicate a significant

relationship between ICW and the test variables. The significant positive relationships

95

between ICW and AU_FEES, LOSS, and MERGER are expected. The significant

negative relationships between ICW and SIZE, QUICK, and INITIAL are also expected.

When evaluating the results using only 2017-2018 control variable years, the model

becomes insignificant and is not tabulated.

TABLE 12 - Regression of Audit Quality on Senior-Level Audit Human Capital at the Audit Office Department (H1b)

Panel A: Using RESTATE as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Variable Wald

Intercept

?

-2.248

3.464

0.063*

-22.539

0.000

0.999

AU_DEPT

-

-0.527

1.855

0.087*

-1.178

0.918

0.169

LG_OFFICE_SEN_LVL

-

-0.242

6.335

0.006***

-0.416

1.918

0.083*

AUDIT_CLIENTS

-

-0.329

4.299

0.019**

-0.138

0.110

0.370

ICW

+

1.403

106.406

0.000***

2.152

30.874

0.000***

AU_FEES

+

0.341

24.735

0.000***

0.289

1.797

0.090*

SIZE

-

-0.171

18.104

0.000***

-0.205

2.581

0.054*

LOSS

+

0.010

0.014

0.453

0.462

2.815

0.093*

EQUITY_MULTIPLIER

+

0.007

3.638

0.028*

0.004

0.131

0.359

QUICK

-

-0.021

1.753

0.093*

-0.046

0.675

0.206

OPSEG

+

0.005

0.392

0.266

0.028

0.217

0.321

GEOSEG

+

0.001

0.037

0.424

0.008

0.033

0.428

LnPOPULATION

-

0.017

0.167

0.683

0.188

2.091

0.148

FIRM_TENURE

+

0.230

3.635

0.029**

0.126

0.114

0.368

NATIONAL_LEADER

-

-0.147

4.079

0.022**

-0.095

0.186

0.333

Expected Sign

Estimate

Wald

p

-

value

Estimate

p

-

value

CITY_LEADER

-

0.010

0.018

0.893

0.214

0.863

0.353

NONAUDIT

+

0.009

0.720

0.198

0.029

0.763

0.191

FEM_AU

?

-1.475

5.849

0.016**

-1.839

0.937

0.333

INITIAL

+

-0.010

0.005

0.945

0.690

1.983

0.080*

MERGER

+

-0.014

0.038

0.845

0.077

0.123

0.363

CHANGE_IN_RECEIVABLES

+

-0.001

0.129

0.719

-0.017

0.013

0.909

CHANGE_IN_INVENTORIES

+

-0.004

0.022

0.881

-0.005

0.008

0.930

CHANGE_IN_CASH_SALES

+

-0.002

1.064

0.302

-0.071

1.739

0.187

Year and Industry fixed effects

Yes

Number of Restatements

1,025

122

Total Number of Observations

11,301

1,524

Model p-value

0.000***

Pseudo R2

5.7%

19.2%

Regression of Audit Quality on Senior-Level Audit Human Capital at the Audit Office Department (H1b)

Panel B: Using ICW as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Variable Wald

Expected Sign

Estimate

Wald

p

-

value

Estimate

p

-

value

Intercept ? -26.593 0.000 0.998 -21.298 0.000 0.999

AU_DEPT + 0.543 0.747 0.194 0.541 0.095 0.379

LG_OFFICE_SEN_LVL + -0.211 1.459 0.227 -0.041 0.008 0.931

AUDIT_CLIENTS + 0.158 0.613 0.217 -0.944 1.031 0.310

AU_FEES + 1.455 143.407 0.000*** 1.243 11.251

0.000***

SIZE - -0.802 125.676 0.000*** -0.792 13.550

0.000***

LOSS + 0.196 1.899 0.084* 0.081 0.045 0.416

EQUITY_MULTIPLIER + 0.006 0.894 0.172 -0.012 0.552 0.457

QUICK - -0.043 1.668 0.099* -0.213 1.990 0.079*

OPSEG + -0.011 0.428 0.513 0.002 0.000 0.493

GEOSEG + 0.001 0.005 0.471 -0.029 0.163 0.686

LnPOPULATION - 0.078 1.129 0.288 -0.134 0.427 0.257

FIRM_TENURE - 0.242 1.439 0.230 0.616 1.406 0.236

NATIONAL_LEADER + -0.109 0.699 0.403 0.099 0.082 0.387

CITY_LEADER + -0.154 1.251 0.263 0.18 0.243 0.311

NONAUDIT + 0.004 0.047 0.414 -0.005 0.011 0.918

FEM_AU ? 1.635 2.483 0.115 3.188 1.022 0.312

INITIAL - -0.334 1.732 0.094* 0.123 0.032 0.857

MERGER + 0.264 4.257 0.020** 0.515 2.238 0.068*

Year and Industry fixed effects Yes Yes

Number of Control Weaknesses 308 45

Total Number of Observations 11,301 1,565

Model p-value 0.000*** 0.129

Pseudo R2 13.3% 23.2%

Regression of Audit Quality on Senior-Level Audit Human Capital at the Audit Office Department (H1b)

Panel C: Using GCM as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Variable Wald p-value Wald

Intercept

?

10.620

14.170

0.000***

7.504

0.000

1.000

AU_DEPT

+

-1.516

2.160

0.142

-5.797

1.892

0.169

LG_OFFICE_SEN_LVL

+

-0.452

2.885

0.089*

-0.976

0.872

0.350

AUDIT_CLIENTS

-

-0.187

0.252

0.308

0.970

0.673

0.412

ICW

+

0.145

0.100

0.376

-15.361

0.000

0.998

AU_FEES

+

0.172

1.623

0.102

0.117

0.033

0.214

SIZE

-

-0.464

24.894

0.000***

-0.780

2.956

0.022**

LOSS

+

0.397

0.574

0.225

18.054

0.000

0.250

EQUITY_MULTIPLIER

+

-0.003

0.120

0.729

-0.025

0.153

0.696

QUICK

-

-0.160

18.018

0.000***

-0.108

1.126

0.073*

OPSEG

-

0.006

0.024

0.878

-0.224

0.428

0.129

Expected Sign

Estimate

p

-

value

GEOSEG

-

-0.048

4.733

0.015**

0.196

2.254

0.133

LnPOPULATION

-

-0.162

2.051

0.076*

-0.707

2.378

0.031**

FIRM_TENURE

-

0.318

1.873

0.171

0.701

0.727

0.394

NATIONAL_LEADER

+

-0.271

2.033

0.154

-0.984

2.126

0.145

CITY_LEADER

+

0.171

0.815

0.184

0.650

0.905

0.086*

NONAUDIT

+

0.002

0.010

0.460

-0.073

1.604

0.205

FEM_AU

?

2.455

2.693

0.101

-3.916

0.473

0.492

INITIAL

-

-0.236

0.835

0.181

-0.896

0.648

0.106

MERGER

-

-0.468

3.120

0.039**

-0.212

0.056

0.204

REPORTLAG

+

0.002

1.385

0.120

0.057

4.337

0.010***

BANKRUPTCY

-

-0.070

35.975

0.000***

-0.076

3.131

0.020**

Year and Industry fixed effects

Yes

Number of Going Concern Modifications

235

19

Total Number of Observations

3,705

481

Model p-value

0.000***

Pseudo R2

40.2%

59.1%

Regression of Audit Quality on Senior-Level Audit Human Capital at the Audit Office Department (H1b)

Panel D: Using AU_FEES as the Dependent Variable

Control Variable Years: 2011-2018

Control Variable Years: 2017-2018

Variable

t-value

p-value

t-value

p-value

Intercept

?

3.991

31.320

0.000***

4.546

13.920

0.000***

AU_DEPT

+

-0.277

-5.261

0.000***

-0.322

-2.383

0.017**

LG_OFFICE_SEN_LVL

+

0.090

6.496

0.000***

0.077

2.184

0.015**

AUDIT_CLIENTS

+

-0.163

-8.527

0.000***

-0.145

-3.129

0.002***

ICW

+

0.344

11.875

0.000***

0.240

3.323

0.000***

RESTATE

+

0.086

5.243

0.000***

0.074

1.639

0.051*

GCM

+

0.116

3.441

0.000***

0.235

2.741

0.003***

SIZE

+

0.461

128.514

0.000***

0.446

47.565

0.000***

LOSS

+

0.165

12.803

0.000***

0.125

3.834

0.000***

EQUITY_MULTIPLIER

+

0.001

0.949

0.171

0.001

0.974

0.165

QUICK

-

-0.026

-13.590

0.000***

-0.025

-4.646

0.000***

OPSEG

+

0.012

9.467

0.000***

0.032

4.341

0.000***

GEOSEG

+

0.008

11.120

0.000***

0.027

5.466

0.000***

lnPOPULATION

+

0.019

3.096

0.001***

0.015

0.987

0.162

FIRM_TENURE

+

0.058

3.323

0.000***

-0.029

-0.664

0.507

NATIONAL_LEADER

+

-0.011

-1.144

0.252

0.008

0.338

0.368

CITY_LEADER

+

0.035

3.318

0.000***

0.037

1.374

0.085*

NONAUDIT

+

0.017

12.458

0.000***

0.017

4.934

0.000***

FEM_AU

?

0.468

5.473

0.000***

0.260

1.189

0.235

INITIAL

-

0.100

4.707

0.000***

0.083

1.314

0.190

FOREIGN

+

0.049

7.899

0.000***

0.017

1.282

0.100*

MERGER

+

0.054

5.157

0.000***

0.062

2.376

0.009***

DCFO

+

0.113

6.529

0.000***

0.125

2.726

0.003***

GROWTH

+

0.013

1.930

0.027**

0.020

0.934

0.176

Year and Industry fixed effects

Yes

Total Number of Observations

11,301

1,565

Model p-value

0.000***

R2

77.2%

78.9%

Expected Sign

Estimate

102

104

Panel C of Table 12 presents the results from the logistic regression of GCM on

the test and control variables for Hypothesis 1b. The results show a significant negative

relationship between GCM and LG_OFFICE_SEN_LVL (p = 0.012), unexpectedly

indicating that offices with more senior-level auditors are less likely to report going

concern opinion modifications. The significant negative relationship between GCM and

BANKRUPTCY is expected. The significant negative relationships between GCM and

SIZE, QUICK, GEOSEG, and MERGER are also expected. When evaluating the results

using only 2017-2018 control variable years, the test variables are not significant, and

none of the control variable results are unexpected.

Panel D of Table 12 presents the results from the OLS regression of AU_FEES on

the test and control variables for Hypothesis 1b. The results indicate a significant

negative relationship between AU_FEES and AU_DEPT (p < 0.01) and a significant

positive relationship between AU_FEES and LG_OFFICE_SEN_LVL (p < 0.01). These

results can potentially be explained in two ways. First, this result might indicate that

offices with more available human capital are more efficient, a finding strengthened by

the significant negative relationship between AU_FEES and AUDIT_CLIENTS (p <

0.01), where offices with more audit clients per auditor are able to create efficiencies

among similar clients to scale fees. Second, the results might indicate that offices with a

larger proportion of senior-level auditors, rather than just more overall auditors, bill

greater amounts due to their expertise. The findings indicate significant positive

relationships between AU_FEES and ICW, RESTATE, SIZE, LOSS, OP_SEG,

GEO_SEG, LnPOPULATION, FIRM_TENURE, CITY_LEADER (when controlling for

105

NATIONAL_LEADER), NONAUDIT, INITIAL, FOREIGN, MERGER, REPORTLAG,

DCFO, and GROWTH, and a significant negative relationship between AU_FEES and

QUICK. In addition, the results continue to indicate the presence of a significant female

audit fee premium when evaluating the positive relationship between AU_FEES and

FEM_AU (p < 0.01). When evaluating the results using only 2017-2018 control variable

years, the findings also indicate a significant negative relationship between AU_FEES

and AU_DEPT and a significant positive relationship between AU_FEES and

LG_OFFICE_SEN_LVL, while none of the control variable results are unexpected.

Results Summary for Senior-Level Audit Human Capital at the Audit Office Department

and Audit Quality (H1b)

The test results for H1b first suggest that audit offices with greater amounts of

senior-level auditors have clients with fewer financial restatements, fewer going concern

opinion modifications, and have no impact on the discovery of material internal control

weaknesses, a mixed conclusion when compared to the expectations of H1b. Taken

together, it might be that offices with larger quantities of senior-level auditors have

“trained” their clients to maintain high quality financial reporting and internal controls,

and have eliminated risky or illiquid clients. It could also be that the interactions

between lower-level auditors and client management are ineffective in reducing the

amount of material internal control weaknesses, which makes sense given that auditor-

client negotiations typically do not occur at lower staffing levels. Finally, the results

show audit fees tend to have a negative relationship to larger overall audit departments,

and tend to have a positive relationship to audit departments with greater proportions of

106

senior-level auditors. It might be that offices with larger proportions of senior-level

auditors cost more due to their level of experience, while at the overall audit department

level, offices with greater amounts of total overall audit human capital are able to create

enough efficiencies to drive down audit cost.

Supplemental Analysis for Senior-Level Audit Human Capital at the Audit Office

Department and Audit Quality (H1b)

In supplemental analysis of H1b testing, the dependent variable RESTATE is again

modified in three ways. First, RESTATE is replaced with RESTATE_NO_ICW. The

model for this test (untabulated) is not significant (p = 0.633). Second, RESTATE was

replaced by RESTATE_CE. The model for this test (untabulated) is significant (p =

0.014), with a pseudo R-square of 0.131; however, the test variables are not significant.

Finally, RESTATE is replaced by SEC_REG_FAR, to test for restatements that occurred at

an auditor office in the same city as one of the eleven SEC regional offices. The model

for this test (untabulated) is significant (p < 0.01), with a pseudo R-square of 0.102, and

produces a significant negative relationship with AU_DEPT and a significant positive

relationship with LG_OFFICE_SEN_LVL. This finding is unexpected, as one would

expect that offices with greater amounts of senior-level human capital and presumably

more experience would be associated with fewer restatements. It is possible that the

nearness to an SEC regional office affects audit offices with greater amounts of

seniorlevel auditors to a greater extent than other audit offices, perhaps because senior-

107

level auditors feel regulatory pressure more strongly than other auditors; therefore this

finding warrants future research.

Education Quality of Audit Human Capital at the Audit Office Department and Audit

Quality (H1c)

Table 13 provides the regression results of the relationship between audit human

capital and audit quality for Hypothesis 1c. Panels A1-A4 of Table 13 present the results

from the logistic regressions of RESTATE on the test and control variables for each

staffing level (A1=partner, A2=principal, A3=senior manager-manager, and A4=senior

associate-associate). The results indicate significant positive associations between

RESTATE and PG_AU_PRIN (p = 0.055), UG_ELITE_AU_SMGR_MGR (p < 0.01), and

UG_MSA_AU_SEN_ASSOC (p = 0.054), and a significant negative association between

RESTATE and UG_AACSB_ACC_AU_PART (p = 0.068). These mixed results at the

partner, principal, senior manager/manager, and senior associate/associate levels are

unexpected, but might indicate that audit departments which are disproportioned (for

example, “top-heavy” and/or “bottom-heavy”) perform poorer than those that are more

proportionately staffed, perhaps because the client workload is not distributed efficiently.

The results also indicate significant positive relationships at all staffing levels between

RESTATE and ICW, AU_FEES, EQUITY_MULTIPLIER and FIRM_TENURE, and

significant negative relationships between RESTATE and AUDIT_CLIENTS, QUICK,

NATIONAL_LEADER, FEM_AU, and SIZE.

TABLE 13 - Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department

(H1c)

Panel A1: Using RESTATE as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Variable p-value Wald

Intercept

?

-2.271

6.917

0.048**

-22.525

0.000

0.999

PG_AU_PART

-

0.181

0.758

0.384

0.307

0.223

0.637

UG_MSA_AU_PART

-

-0.162

0.729

0.197

-0.208

0.121

0.364

UG_ELITE_AU_PART

-

-0.199

0.781

0.189

-0.586

0.572

0.225

UG_AACSB_ACC_AU_PART

-

-0.314

2.220

0.068*

-1.612

4.988

0.013**

AUDIT_CLIENTS

-

-0.347

4.526

0.017**

-0.176

0.163

0.343

ICW

+

1.406

107.028

0.000***

2.196

31.833

0.000***

AU_FEES

+

0.332

23.799

0.000***

0.283

1.716

0.095*

SIZE

-

-0.167

17.178

0.000***

-0.189

2.200

0.069*

LOSS

+

0.002

0.001

0.490

-0.494

3.197

0.074*

EQUITY_MULTIPLIER

+

0.007

3.748

0.027**

0.005

0.155

0.347

QUICK

-

-0.023

2.107

0.074*

-0.045

0.622

0.215

OPSEG

+

0.004

0.296

0.293

0.018

0.090

0.382

GEOSEG

+

0.001

0.030

0.431

0.004

0.009

0.463

Expected Sign

Estimate

Wald

Estimate

p

-

value

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

Pseudo R2

LnPOPULATION

-

0.004

0.011

0.915

0.164

1.641

0.200

FIRM_TENURE

+

0.240

3.958

0.024**

0.194

0.268

0.303

NATIONAL_LEADER

-

-0.148

4.065

0.022**

-0.112

0.250

0.309

CITY_LEADER

-

0.021

0.073

0.787

0.236

1.013

0.314

NONAUDIT

+

0.009

0.724

0.198

0.032

0.920

0.169

FEM_AU

?

-1.731

8.041

0.005***

-2.960

2.439

0.118

INITIAL

+

-0.010

0.004

0.948

0.692

1.982

0.080*

MERGER

+

-0.015

0.041

0.840

0.110

0.252

0.308

CHANGE_IN_RECEIVABLES

+

-0.001

0.112

0.738

-0.003

0.000

0.983

CHANGE_IN_INVENTORIES

+

-0.005

0.027

0.869

-0.005

0.009

0.925

CHANGE_IN_CASH_SALES

+

-0.002

1.032

0.310

-0.070

1.670

0.196

Year and Industry fixed effects

Yes

Number of Restatements

1,025

122

Total Number of Observations

11,301

1,524

Model p-value

0.000***

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

Pseudo R2

5.7%

19.9%

Panel A2: Using RESTATE as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Variable Wald

Intercept

?

-2.165

3.558

0.059*

-22.772

0.000

0.999

PG_AU_PRIN

-

0.279

3.695

0.055*

0.386

0.803

0.370

UG_MSA_AU_PRIN

-

-0.018

0.013

0.456

-0.596

1.381

0.120

UG_ELITE_AU_PRIN

-

0.123

0.294

0.588

-0.447

0.363

0.274

UG_AACSB_ACC_AU_PRIN

-

-0.177

0.942

0.166

-0.346

0.427

0.257

AUDIT_CLIENTS

-

-0.317

3.836

0.025**

0.012

0.001

0.978

ICW

+

1.412

107.634

0.000***

2.104

29.196

0.000***

AU_FEES

+

0.322

22.139

0.000***

0.273

1.609

0.051*

SIZE

-

-0.165

16.764

0.000***

-0.199

2.452

0.059*

LOSS

+

0.000

0.499

-0.491

3.165

0.038**

EQUITY_MULTIPLIER

+

0.007

3.864

0.025**

0.005

0.147

0.351

Expected Sign

Estimate

Wald

p

-

value

Estimate

p

-

value

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

Pseudo R2

QUICK

-

-0.023

2.166

0.071*

-0.047

0.702

0.201

OPSEG

+

0.005

0.432

0.256

0.024

0.156

0.347

GEOSEG

+

0.001

0.041

0.420

0.008

0.039

0.422

LnPOPULATION

-

-0.005

0.016

0.450

0.186

2.159

0.142

FIRM_TENURE

+

0.228

3.588

0.029**

0.122

0.108

0.372

NATIONAL_LEADER

-

-0.142

3.759

0.027**

-0.103

0.219

0.320

CITY_LEADER

-

0.022

0.084

0.772

0.184

0.634

0.426

NONAUDIT

+

0.009

0.808

0.185

0.030

0.836

0.180

FEM_AU

?

-1.587

6.581

0.010***

-2.263

1.424

0.233

INITIAL

+

-0.009

0.004

0.951

0.717

2.162

0.071*

MERGER

+

-0.014

0.038

0.845

0.097

0.197

0.329

CHANGE_IN_RECEIVABLES

+

-0.001

0.128

0.720

-0.017

0.014

0.905

CHANGE_IN_INVENTORIES

+

-0.004

0.018

0.893

-0.002

0.002

0.963

CHANGE_IN_CASH_SALES

+

-0.002

1.020

0.313

-0.072

1.751

0.186

Year and Industry fixed effects

Yes

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

Pseudo R2

Number of Restatements

1,025

122

Total Number of Observations

11,301

1,524

Model p-value

0.000***

5.7%

19.3%

Panel A3: Using RESTATE as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Variable p-value Wald

Intercept

?

-2.407

4.447

0.035**

-22.718

0.000

0.999

PG_AU_SMGR_MGR

-

-0.277

1.614

0.102

0.178

0.070

0.791

UG_MSA_AU_SMGR_MGR

-

0.301

1.947

0.163

-0.274

0.172

0.678

UG_ELITE_AU_SMGR_MGR

-

1.187

12.618

0.000***

3.500

15.988

0.000***

UG_AACSB_ACC_AU_SMGR_MGR

-

0.113

0.172

0.678

-0.619

0.504

0.478

AUDIT_CLIENTS

-

-0.383

5.787

0.008***

-0.348

0.707

0.201

ICW

+

1.405

106.241

0.000***

2.220

32.340

0.000***

AU_FEES

+

0.342

24.734

0.000***

0.335

2.333

0.064*

Expected Sign

Estimate

Wald

Estimate

p

-

value

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

Pseudo R2

SIZE

-

-0.172

18.305

0.000***

-0.254

3.891

0.025**

LOSS

+

0.000

0.498

-0.524

3.589

0.058*

EQUITY_MULTIPLIER

+

0.007

3.847

0.025**

0.004

0.136

0.357

QUICK

-

-0.022

1.973

0.080*

-0.049

0.781

0.189

OPSEG

+

0.005

0.356

0.276

0.029

0.240

0.312

GEOSEG

+

0.001

0.035

0.426

0.004

0.011

0.458

LnPOPULATION

-

-0.015

0.132

0.359

0.138

1.111

0.292

FIRM_TENURE

+

0.239

3.929

0.024**

0.164

0.193

0.330

NATIONAL_LEADER

-

-0.144

3.902

0.024**

-0.108

0.236

0.314

CITY_LEADER

-

0.010

0.017

0.896

0.233

0.997

0.318

NONAUDIT

+

0.011

1.038

0.154

0.035

1.067

0.151

FEM_AU

?

-1.017

2.759

0.097*

-0.841

0.191

0.662

INITIAL

+

-0.015

0.010

0.921

0.714

2.135

0.072*

MERGER

+

-0.022

0.087

0.768

0.033

0.023

0.441

CHANGE_IN_RECEIVABLES

+

-0.001

0.146

0.703

-0.015

0.010

0.920

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

Pseudo R2

CHANGE_IN_INVENTORIES

+

-0.004

0.023

0.879

-0.008

0.017

0.896

CHANGE_IN_CASH_SALES

+

-0.002

0.892

0.345

-0.066

1.505

0.220

Year and Industry fixed effects

Yes

Number of Restatements

1,025

122

Total Number of Observations

11,301

1,524

Model p-value

0.000***

5.9%

21.0%

Panel A4: Using RESTATE as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Variable p-value Wald

Intercept

?

-2.839

5.994

0.014**

-24.064

0.000

0.999

PG_AU_SEN_ASSOC

-

0.376

1.361

0.243

0.014

0.000

0.988

UG_MSA_AU_SEN_ASSOC

-

0.493

3.710

0.054*

1.425

3.134

0.077*

UG_ELITE_AU_SEN_ASSOC

-

0.188

0.199

0.656

-0.368

0.081

0.388

UG_AACSB_ACC_AU_SEN_ASSOC

-

0.353

1.186

0.276

0.184

0.035

0.852

Expected Sign

Estimate

Wald

Estimate

p

-

value

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

Pseudo R2

AUDIT_CLIENTS

-

-0.300

3.505

0.031**

-0.097

0.048

0.414

ICW

+

1.413

107.514

0.000***

2.171

31.098

0.000***

AU_FEES

+

0.340

24.772

0.000***

0.299

1.934

0.082*

SIZE

-

-0.167

17.198

0.000***

-0.196

2.348

0.063*

LOSS

+

0.009

0.011

0.458

-0.487

3.121

0.077*

EQUITY_MULTIPLIER

+

0.007

4.052

0.022**

0.005

0.163

0.343

QUICK

-

-0.022

2.034

0.077*

-0.049

0.747

0.194

OPSEG

+

0.005

0.355

0.276

0.029

0.239

0.313

GEOSEG

+

0.001

0.012

0.457

0.005

0.011

0.476

LnPOPULATION

-

-0.004

0.008

0.466

0.185

2.051

0.152

FIRM_TENURE

+

0.231

3.681

0.028**

0.132

0.125

0.362

NATIONAL_LEADER

-

-0.159

4.681

0.016**

-0.099

0.201

0.327

CITY_LEADER

-

-0.006

0.007

0.467

0.167

0.525

0.469

NONAUDIT

+

0.010

0.964

0.163

0.030

0.792

0.187

FEM_AU

?

-1.326

4.815

0.028**

-1.715

0.900

0.343

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

Pseudo R2

INITIAL

+

-0.009

0.004

0.952

0.735

2.238

0.068*

MERGER

+

-0.016

0.048

0.827

0.091

0.170

0.340

CHANGE_IN_RECEIVABLES

+

-0.001

0.130

0.718

-0.011

0.008

0.931

CHANGE_IN_INVENTORIES

+

-0.004

0.017

0.896

-0.005

0.009

0.924

CHANGE_IN_CASH_SALES

+

-0.002

0.987

0.321

-0.072

1.786

0.181

Year and Industry fixed effects

Yes

Number of Restatements

1,025

122

Total Number of Observations

11,301

1,524

Model p-value

0.000***

5.8%

19.4%

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

Panel B1: Using ICW as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Variable Wald

Intercept ? -25.745 0.000 0.998 -19.279 0.000 0.999

PG_AU_PART + -0.437 1.119 0.290 -0.548 0.276 0.599

UG_MSA_AU_PART + -0.136 0.139 0.709 0.334 0.120 0.365

UG_ELITE_AU_PART + -0.171 0.147 0.702 -1.217 0.887 0.346

UG_AACSB_ACC_AU_PART + -0.495 1.491 0.222 0.851 0.853 0.178

AUDIT_CLIENTS + 0.121 0.335 0.282 -0.975 1.021 0.312

AU_FEES + 1.433 140.821 0.000*** 1.239 11.243

0.000***

SIZE - -0.792 122.962 0.000*** -0.792 13.283

0.000***

LOSS + 0.180 1.601 0.103 0.050 0.017 0.448

EQUITY_MULTIPLIER + 0.006 0.830 0.181 -0.009 0.353 0.552

QUICK - -0.046 1.839 0.088* -0.222 2.131 0.072*

OPSEG + -0.010 0.364 0.546 0.016 0.014 0.453

GEOSEG + 0.000 0.001 0.491 -0.034 0.220 0.639

LnPOPULATION - 0.053 0.539 0.463 -0.095 0.207 0.325

FIRM_TENURE - 0.244 1.456 0.228 0.584 1.264 0.261

NATIONAL_LEADER + -0.104 0.628 0.428 0.122 0.124 0.363

CITY_LEADER + -0.137 0.990 0.320 0.149 0.164 0.343

NONAUDIT + 0.004 0.045 0.416 -0.006 0.017 0.896

Expected Sign

Estimate

Wald

p

-

value

Estimate

p

-

value

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

FEM_AU ? 1.191 1.189 0.275 3.738 1.303 0.254

INITIAL - -0.332 1.707 0.096* 0.177 0.066 0.797

MERGER + 0.267 4.353 0.019** 0.537 2.427 0.060*

Year and Industry fixed effects Yes Yes

Number of Control Weaknesses 308 45

Total Number of Observations 11,301 1,565

Model p-value 0.000*** 0.137

Pseudo R2 13.3% 23.7%

Panel B2: Using ICW as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Variable Wald

Intercept

?

-26.047

0.000

0.998

-21.356

0.000

0.999

PG_AU_PRIN

+

-0.111

0.150

0.699

0.368

0.260

0.305

UG_MSA_AU_PRIN

+

-0.368

1.400

0.237

-0.163

0.043

0.835

UG_ELITE_AU_PRIN

+

-0.614

1.650

0.199

-2.313

1.634

0.201

UG_AACSB_ACC_AU_PRIN

+

-0.154

0.171

0.679

-1.427

1.403

0.236

Expected Sign

Estimate

Wald

p

-

value

Estimate

p

-

value

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

AUDIT_CLIENTS

+

0.220

1.080

0.150

-0.928

0.823

0.364

AU_FEES

+

1.430

139.905

0.000***

1.200

10.435

0.000***

SIZE

-

-0.791

122.124

0.000***

-0.761

12.522

0.000***

LOSS

+

0.204

2.049

0.076*

0.072

0.035

0.426

EQUITY_MULTIPLIER

+

0.006

0.939

0.166

-0.012

0.577

0.448

QUICK

-

-0.046

1.889

0.085*

-0.212

2.006

0.079*

OPSEG

+

-0.009

0.314

0.575

0.015

0.013

0.454

GEOSEG

+

0.001

0.022

0.442

-0.029

0.168

0.682

LnPOPULATION

-

0.067

0.863

0.353

-0.084

0.169

0.341

FIRM_TENURE

-

0.242

1.437

0.231

0.697

1.794

0.180

NATIONAL_LEADER

+

-0.111

0.721

0.396

0.125

0.130

0.359

CITY_LEADER

+

-0.168

1.467

0.226

0.138

0.142

0.354

NONAUDIT

+

0.002

0.010

0.461

-0.014

0.086

0.770

FEM_AU

?

1.224

1.268

0.260

2.350

0.541

0.462

INITIAL

-

-0.336

1.760

0.093*

0.037

0.003

0.958

MERGER

+

0.263

4.220

0.020**

0.505

2.123

0.073*

Year and Industry fixed effects

Yes

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

Number of Control Weaknesses

308

45

Total Number of Observations

11,301

1,565

Model p-value

0.000***

0.097*

Pseudo R2

13.4%

24.4%

Panel B3: Using ICW as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Variable Wald Wald

Intercept ? -25.686 0.000 0.998 -20.063 0.000 0.999

PG_AU_SMGR_MGR + 0.599 2.429 0.060* 1.292 1.897 0.084*

UG_MSA_AU_SMGR_MGR + 0.774 3.989 0.023** 0.268 0.066 0.399

UG_ELITE_AU_SMGR_MGR + 0.478 0.524 0.235 -1.139 0.267 0.605

UG_AACSB_ACC_AU_SMGR_MGR + -0.428 0.722 0.396 -1.470 1.074 0.300

AUDIT_CLIENTS + 0.092 0.181 0.335 -1.280 1.343 0.246

AU_FEES + 1.485 147.268 0.000*** 1.295 11.981

0.000***

SIZE - -0.806 126.763 0.000*** -0.817 13.977

0.000***

LOSS + 0.211 2.180 0.070* 0.160 0.174 0.338

Expected Sign

Estimate

p

-

value

Estimate

p

-

value

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

EQUITY_MULTIPLIER + 0.006 0.811 0.184 -0.013 0.656 0.418

QUICK - -0.043 1.642 0.100* -0.203 1.760 0.093*

OPSEG + -0.008 0.248 0.618 0.006 0.002 0.483

GEOSEG + -0.001 0.006 0.937 -0.023 0.109 0.741

LnPOPULATION - -0.020 0.075 0.392 -0.209 0.937 0.167

FIRM_TENURE - 0.240 1.398 0.237 0.630 1.474 0.225

NATIONAL_LEADER + -0.091 0.485 0.486 0.074 0.045 0.416

CITY_LEADER + -0.179 1.682 0.195 0.115 0.095 0.379

NONAUDIT + 0.001 0.003 0.477 -0.005 0.011 0.459

FEM_AU ? 1.454 1.875 0.171 2.302 0.559 0.228

INITIAL - -0.345 1.818 0.089* 0.074 0.011 0.458

MERGER + 0.248 3.764 0.026** 0.528 2.299 0.065*

Year and Industry fixed effects Yes Yes

Number of Control Weaknesses 308 45

Total Number of Observations 11,301 1,565

Model p-value 0.000*** 0.114

Pseudo R2 13.6% 24.1%

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

Panel B4: Using ICW as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Variable Wald

Intercept ? -25.856 0.000 0.998 -20.941 0.000 0.999

PG_AU_SEN_ASSOC + -0.451 0.583 0.445 -0.198 0.017 0.895

UG_MSA_AU_SEN_ASSOC + 0.543 1.330 0.125 -0.464 0.132 0.716

UG_ELITE_AU_SEN_ASSOC + -0.963 1.443 0.230 -3.211 1.239 0.266

UG_AACSB_ACC_AU_SEN_ASSOC + -0.598 0.914 0.339 -1.292 0.597 0.440

AUDIT_CLIENTS + 0.216 1.019 0.157 -0.984 0.898 0.343

AU_FEES + 1.430 140.391 0.000*** 1.244 11.421

0.000***

SIZE - -0.786 121.080 0.000*** -0.793 13.389

0.000***

LOSS + 0.202 1.991 0.079* 0.074 0.038 0.423

EQUITY_MULTIPLIER + 0.006 0.898 0.172 -0.012 0.551 0.458

QUICK - -0.048 2.048 0.076* -0.211 1.912 0.084*

OPSEG + -0.009 0.272 0.602 0.010 0.006 0.471

GEOSEG + 0.001 0.024 0.439 -0.029 0.164 0.686

LnPOPULATION - 0.054 0.553 0.457 -0.099 0.221 0.319

FIRM_TENURE - 0.239 1.404 0.236 0.657 1.613 0.204

NATIONAL_LEADER + -0.088 0.455 0.500 0.136 0.156 0.347

CITY_LEADER + -0.156 1.278 0.258 0.203 0.306 0.290

NONAUDIT + 0.001 0.002 0.481 -0.008 0.031 0.861

FEM_AU ? 1.285 1.322 0.250 2.712 0.671 0.413

Expected Sign

Estimate

Wald

p

-

value

Estimate

p

-

value

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

INITIAL - -0.333 1.729 0.095* 0.089 0.017 0.896

MERGER + 0.259 4.124 0.021** 0.527 2.346 0.063*

Year and Industry fixed effects Yes Yes

Number of Control Weaknesses 308 45

Total Number of Observations 11,301 1,565

Model p-value 0.000*** 0.121

Pseudo R2 13.4% 23.9%

Panel C1: Using GCM as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Variable Wald p-value Wald

Intercept ? 10.830 15.011 0.000*** 7.300 0.000 1.000

PG_AU_PART + -0.052 0.008 0.930 -4.049 2.689 0.101

UG_MSA_AU_PART + -0.408 0.668 0.414 -2.416 2.205 0.138

UG_ELITE_AU_PART + 0.480 0.577 0.224 1.161 0.392 0.266

UG_AACSB_ACC_AU_PART + -1.081 3.455 0.063* -0.164 0.007 0.934

AUDIT_CLIENTS - -0.349 0.855 0.178 1.207 1.013 0.314

ICW + 0.112 0.059 0.404 -15.375 0.000 0.998

AU_FEES + 0.171 1.550 0.107 0.317 0.227 0.317

Expected Sign

Estimate

p

-

value

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

SIZE - -0.468 24.716 0.000*** -1.021 4.457

0.018**

LOSS + 0.386 0.540 0.232 17.655 0.000 0.499

EQUITY_MULTIPLIER + -0.004 0.233 0.629 -0.017 0.082 0.774

QUICK - -0.164 18.921 0.000*** -0.106 1.107 0.147

OPSEG - 0.003 0.005 0.944 -0.198 0.299 0.292

GEOSEG - -0.053 5.526 0.001*** 0.243 3.202

0.074*

LnPOPULATION - -0.196 2.991 0.042** -0.674 2.721

0.050**

FIRM_TENURE - 0.334 2.080 0.149 1.055 1.551 0.213

NATIONAL_LEADER + -0.313 2.632 0.105 -1.018 1.984 0.159

CITY_LEADER + 0.174 0.831 0.181 0.322 0.221 0.319

NONAUDIT + 0.002 0.016 0.450 -0.065 1.181 0.277

FEM_AU ? 2.035 1.958 0.162 -2.042 0.146 0.703

INITIAL - -0.224 0.745 0.194 -1.229 1.042 0.154

MERGER - -0.441 2.778 0.048** 0.022 0.001 0.980

REPORTLAG + 0.002 1.398 0.119 0.057 4.184

0.021**

BANKRUPTCY - -0.067 33.148 0.000*** -0.072 2.656

0.052*

Year and Industry fixed effects Yes Yes

Number of Going Concern Modifications 235 19

Total Number of Observations 3,705 481

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

Model p-value 0.000*** 0.000***

Pseudo R2 40.2% 60.0%

Panel C2: Using GCM as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Variable Wald p-value Wald

Intercept

?

10.045

13.999

0.000***

11.337

0.000

1.000

PG_AU_PRIN

+

-0.284

0.457

0.499

-0.730

0.267

0.605

UG_MSA_AU_PRIN

+

-0.022

0.002

0.961

-2.829

2.975

0.085*

UG_ELITE_AU_PRIN

+

0.814

1.640

0.100*

-0.252

0.016

0.901

UG_AACSB_ACC_AU_PRIN

+

-0.043

0.006

0.937

-0.911

0.290

0.591

AUDIT_CLIENTS

-

-0.290

0.579

0.224

1.410

1.091

0.296

ICW

+

0.120

0.067

0.398

-14.888

0.000

0.998

AU_FEES

+

0.175

1.652

0.100*

0.575

0.648

0.211

SIZE

-

-0.455

23.697

0.000***

-0.911

3.719

0.027**

LOSS

+

0.401

0.589

0.222

18.373

0.000

0.499

EQUITY_MULTIPLIER

+

-0.003

0.124

0.725

-0.036

0.299

0.584

QUICK

-

-0.160

18.431

0.000***

-0.116

1.295

0.128

OPSEG

-

0.010

0.073

0.787

0.064

0.032

0.857

Expected Sign

Estimate

p

-

value

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

GEOSEG

-

-0.052

5.568

0.009***

0.224

2.987

0.084*

LnPOPULATION

-

-0.186

2.689

0.051*

-1.079

5.511

0.001***

FIRM_TENURE

-

0.330

2.014

0.156

0.483

0.326

0.568

NATIONAL_LEADER

+

-0.306

2.552

0.110

-1.086

2.286

0.131

CITY_LEADER

+

0.126

0.435

0.255

0.168

0.056

0.407

NONAUDIT

+

-0.001

0.001

0.976

-0.079

1.851

0.174

FEM_AU

?

2.239

2.420

0.120

-6.934

1.787

0.181

INITIAL

-

-0.216

0.696

0.202

-0.589

0.272

0.301

MERGER

-

-0.444

2.815

0.047**

-0.446

0.222

0.319

REPORTLAG

+

0.002

1.412

0.118

0.064

4.546

0.017**

BANKRUPTCY

-

-0.069

35.073

0.000***

-0.085

3.956

0.024**

Year and Industry fixed effects

Yes

Number of Going Concern Modifications

235

19

Total Number of Observations

3,705

481

Model p-value

0.000***

Pseudo R2

40.1%

60.4%

Panel C3: Using GCM as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

Variable Wald p-value Wald

Intercept

?

9.454

11.917

0.001***

3.359

0.000

1.000

PG_AU_SMGR_MGR

+

1.771

10.254

0.000***

1.763

0.865

0.176

UG_MSA_AU_SMGR_MGR

+

0.086

0.026

0.437

-0.113

0.003

0.955

UG_ELITE_AU_SMGR_MGR

+

0.131

0.015

0.452

1.781

0.208

0.325

UG_AACSB_ACC_AU_SMGR_MGR

+

1.055

2.024

0.078*

0.782

0.079

0.390

AUDIT_CLIENTS

-

-0.394

0.983

0.161

1.071

0.789

0.374

ICW

+

0.061

0.017

0.448

-15.817

0.000

0.998

AU_FEES

+

0.241

3.053

0.041**

0.199

0.090

0.382

SIZE

-

-0.469

24.628

0.000***

-0.688

2.260

0.067*

LOSS

+

0.459

0.761

0.192

17.672

0.000

0.499

EQUITY_MULTIPLIER

+

-0.003

0.073

0.787

-0.048

0.538

0.463

QUICK

-

-0.156

17.350

0.000***

-0.092

0.860

0.177

OPSEG

-

0.009

0.056

0.812

-0.225

0.356

0.276

GEOSEG

-

-0.055

5.959

0.008***

0.132

0.805

0.370

LnPOPULATION

-

-0.265

4.994

0.013**

-0.876

3.986

0.023**

FIRM_TENURE

-

0.308

1.720

0.190

0.953

1.416

0.234

Expected Sign

Estimate

p

-

value

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

NATIONAL_LEADER

+

-0.195

1.038

0.308

-0.838

1.376

0.241

CITY_LEADER

+

0.102

0.296

0.293

0.349

0.266

0.303

NONAUDIT

+

-0.003

0.027

0.869

-0.068

1.402

0.236

FEM_AU

?

3.285

5.023

0.025**

-1.269

0.063

0.803

INITIAL

-

-0.225

0.750

0.194

-1.060

0.889

0.173

MERGER

-

-0.454

2.891

0.045**

-0.264

0.080

0.389

REPORTLAG

+

0.003

2.354

0.063*

0.053

3.668

0.028**

BANKRUPTCY

-

-0.071

36.101

0.000***

-0.085

3.727

0.027**

Year and Industry fixed effects

Yes

Number of Going Concern Modifications

235

19

Total Number of Observations

3,705

481

Model p-value

0.000***

Pseudo R2

41.1%

58.9%

Panel C4: Using GCM as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Variable Wald p-value Wald

Intercept

?

9.873

12.787

0.000***

8.664

0.000

1.000

PG_AU_SEN_ASSOC

+

1.605

2.990

0.042**

-1.443

0.279

0.597

UG_MSA_AU_SEN_ASSOC

+

-0.314

0.210

0.647

-1.595

0.436

0.509

UG_ELITE_AU_SEN_ASSOC

+

0.967

0.509

0.238

-1.191

0.055

0.815

UG_AACSB_ACC_AU_SEN_ASSOC

+

-1.185

1.303

0.254

-2.108

0.499

0.480

AUDIT_CLIENTS

-

-0.410

1.116

0.146

0.860

0.479

0.489

ICW

+

0.077

0.028

0.434

-15.736

0.000

0.998

AU_FEES

+

0.189

1.907

0.084**

0.287

0.183

0.335

SIZE

-

-0.458

24.133

0.000***

-0.852

3.208

0.037**

LOSS

+

0.354

0.460

0.249

18.671

0.000

0.499

EQUITY_MULTIPLIER

+

-0.004

0.206

0.650

-0.030

0.189

0.664

QUICK

-

-0.168

19.542

0.000***

-0.104

1.061

0.152

OPSEG

-

0.009

0.061

0.805

-0.007

0.000

0.493

GEOSEG

-

-0.051

5.212

0.011**

0.233

3.044

0.081*

LnPOPULATION

-

-0.185

2.592

0.054*

-0.851

3.375

0.033**

FIRM_TENURE

-

0.336

2.090

0.148

0.712

0.758

0.384

Expected Sign

Estimate

p

-

value

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

NATIONAL_LEADER

+

-0.326

2.936

0.087*

-1.145

2.740

0.098*

CITY_LEADER

+

0.124

0.426

0.257

0.425

0.402

0.263

NONAUDIT

+

0.002

0.013

0.454

-0.061

1.174

0.279

FEM_AU

?

1.355

0.804

0.370

-5.894

1.216

0.270

INITIAL

-

-0.221

0.733

0.196

-0.565

0.252

0.308

MERGER

-

-0.453

2.908

0.044**

-0.487

0.257

0.306

REPORTLAG

+

0.002

1.773

0.092*

0.062

4.234

0.020**

BANKRUPTCY

-

-0.067

33.032

0.000***

-0.076

3.120

0.039**

Year and Industry fixed effects

Yes

Number of Going Concern Modifications

235

19

Total Number of Observations

3,705

481

Model p-value

0.000***

Pseudo R2

40.2%

59.1%

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

Panel D1: Using AU_FEES as the Dependent Variable

Control Variable Years: 2011-2018

Control Variable Years: 2017-2018

Variable

t-value

p-value

t-value

p-value

Intercept

?

3.590

30.190

0.000***

4.100

13.510

0.000***

PG_AU_PART

+

0.021

0.713

0.238

0.061

0.835

0.202

UG_MSA_AU_PART

+

-0.022

-0.880

0.190

0.013

0.195

0.423

UG_ELITE_AU_PART

+

0.012

0.427

0.335

0.044

0.595

0.276

UG_AACSB_ACC_AU_PART

+

-0.023

-0.833

0.203

-0.070

-1.005

0.158

AUDIT_CLIENTS

+

-0.172

-8.821

0.000***

-0.150

-3.166

0.001***

ICW

+

0.342

11.751

0.000***

0.244

3.359

0.000***

RESTATE

+

0.084

5.136

0.000***

0.072

1.587

0.057*

GCM

+

0.123

3.651

0.000***

0.245

2.842

0.003***

SIZE

+

0.464

128.883

0.000***

0.450

47.760

0.000***

LOSS

+

0.171

13.165

0.000***

0.133

4.072

0.000***

EQUITY_MULTIPLIER

+

0.001

0.913

0.181

0.001

0.853

0.197

QUICK

-

-0.025

-13.170

0.000***

-0.024

-4.518

0.000***

OPSEG

+

0.011

9.074

0.000***

0.031

4.134

0.000***

GEOSEG

+

0.008

11.037

0.000***

0.026

5.343

0.000***

lnPOPULATION

+

0.038

6.379

0.000***

0.032

2.142

0.016**

FIRM_TENURE

+

0.061

3.441

0.000***

-0.021

-0.476

0.317

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

NATIONAL_LEADER

+

-0.016

-1.572

0.058*

0.001

0.056

0.478

CITY_LEADER

+

0.026

2.430

0.008***

0.034

1.262

0.104

NONAUDIT

+

0.017

12.610

0.000***

0.016

4.844

0.000***

FEM_AU

?

0.626

7.325

0.000***

0.450

2.081

0.019**

INITIAL

-

0.105

4.907

0.000***

0.081

1.272

0.102

FOREIGN

+

0.050

7.984

0.000***

0.017

1.332

0.092*

MERGER

+

0.053

5.099

0.000***

0.060

2.300

0.011**

DCFO

+

0.117

6.730

0.000***

0.124

2.679

0.004***

GROWTH

+

0.013

1.959

0.025**

0.019

0.872

0.192

Year and Industry fixed effects

Yes

Total Number of Observations

11,301

1,565

Model p-value

0.000***

R2

77.0%

78.7%

Expected Sign

Estimate

Panel D2: Using AU_FEES as the Dependent Variable

Control Variable Years: 2011-2018

Control Variable Years: 2017-2018

Variable

t-value

p-value

t-value

p-value

Intercept

?

3.597

30.105

0.000***

4.123

13.523

0.000***

PG_AU_PRIN

+

0.154

7.393

0.000***

0.194

3.791

0.000***

UG_MSA_AU_PRIN

+

0.046

2.030

0.021**

-0.013

-0.230

0.818

UG_ELITE_AU_PRIN

+

-0.061

-1.932

0.053*

-0.039

-0.496

0.620

UG_AACSB_ACC_AU_PRIN

+

-0.049

-1.979

0.048**

-0.051

-0.857

0.392

AUDIT_CLIENTS

+

-0.162

-8.199

0.000***

-0.130

-2.701

0.007***

ICW

+

0.342

11.781

0.000***

0.236

3.272

0.000***

RESTATE

+

0.082

4.978

0.000***

0.069

1.517

0.065*

GCM

+

0.125

3.708

0.000***

0.248

2.891

0.002***

SIZE

+

0.462

128.465

0.000***

0.446

47.493

0.000***

LOSS

+

0.167

12.898

0.000***

0.132

4.040

0.000***

EQUITY_MULTIPLIER

+

0.001

0.998

0.159

0.001

0.980

0.164

QUICK

-

-0.026

-13.291

0.000***

-0.024

-4.415

0.000***

OPSEG

+

0.011

9.142

0.000***

0.031

4.084

0.000***

GEOSEG

+

0.008

11.015

0.000***

0.027

5.461

0.000***

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

lnPOPULATION

+

0.036

6.058

0.000***

0.031

2.095

0.018**

FIRM_TENURE

+

0.058

3.310

0.000***

-0.028

-0.653

0.514

NATIONAL_LEADER

+

-0.014

-1.375

0.169

0.006

0.235

0.407

CITY_LEADER

+

0.032

3.004

0.002***

0.034

1.276

0.101

NONAUDIT

+

0.017

12.844

0.000***

0.017

5.012

0.000***

FEM_AU

?

0.641

7.425

0.000***

0.483

2.208

0.027**

INITIAL

-

0.102

4.778

0.000***

0.082

1.290

0.197

FOREIGN

+

0.049

7.891

0.000***

0.017

1.333

0.092*

MERGER

+

0.054

5.155

0.000***

0.062

2.383

0.009***

DCFO

+

0.110

6.374

0.000***

0.115

2.487

0.007***

GROWTH

+

0.014

1.986

0.024**

0.018

0.834

0.202

Year and Industry fixed effects

Yes

Total Number of Observations

11,301

1,565

Model p-value

0.000***

R2

77.1%

78.9%

Expected Sign

Estimate

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

Panel D3: Using AU_FEES as the Dependent Variable

Control Variable Years: 2011-2018

Control Variable Years: 2017-2018

Variable

t-value

p-value

t-value

p-value

Intercept

?

3.519

29.961

0.000***

4.020

13.449

0.000***

PG_AU_SMGR_MGR

+

-0.240

-8.102

0.000***

-0.245

-3.348

0.001***

UG_MSA_AU_SMGR_MGR

+

-0.203

-6.733

0.000***

-0.244

-3.246

0.001***

UG_ELITE_AU_SMGR_MGR

+

-0.136

-2.614

0.009***

-0.225

-1.789

0.074*

UG_AACSB_ACC_AU_SMGR_MGR

+

0.072

1.852

0.032**

0.088

0.896

0.185

AUDIT_CLIENTS

+

-0.134

-6.863

0.000***

-0.097

-2.057

0.040**

ICW

+

0.348

12.031

0.000***

0.241

3.357

0.000***

RESTATE

+

0.085

5.239

0.000***

0.080

1.782

0.038**

GCM

+

0.128

3.815

0.000***

0.257

3.015

0.002***

SIZE

+

0.461

128.496

0.000***

0.445

47.667

0.000***

LOSS

+

0.161

12.523

0.000***

0.122

3.760

0.000***

EQUITY_MULTIPLIER

+

0.000

0.750

0.227

0.001

0.888

0.188

QUICK

-

-0.026

-13.626

0.000***

-0.024

-4.552

0.000***

OPSEG

+

0.011

9.230

0.000***

0.031

4.194

0.000***

GEOSEG

+

0.008

11.222

0.000***

0.027

5.557

0.000***

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

LnPOPULATION

+

0.057

9.616

0.000***

0.055

3.671

0.000***

FIRM_TENURE

+

0.060

3.435

0.000***

-0.024

-0.569

0.569

NATIONAL_LEADER

+

-0.020

-2.043

0.041**

0.005

0.180

0.429

CITY_LEADER

+

0.039

3.693

0.000***

0.041

1.529

0.064*

NONAUDIT

+

0.017

12.755

0.000***

0.016

4.845

0.000***

FEM_AU

?

0.570

6.652

0.000***

0.415

1.918

0.055*

INITIAL

-

0.100

4.713

0.000***

0.099

1.570

0.117

FOREIGN

+

0.049

7.910

0.000***

0.016

1.278

0.101

MERGER

+

0.056

5.392

0.000***

0.067

2.599

0.005***

DCFO

+

0.108

6.274

0.000***

0.112

2.453

0.007***

GROWTH

+

0.013

1.964

0.025**

0.016

0.764

0.223

Year and Industry fixed effects

Yes

Total Number of Observations

11,301

1,565

Model p-value

0.000***

R2

77.3%

79.1%

Expected Sign

Estimate

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

Panel D4: Using AU_FEES as the Dependent Variable

Control Variable Years: 2011-2018

Control Variable Years: 2017-2018

Variable

t-value

p-value

t-value

p-value

Intercept

?

3.592

29.694

0.000***

4.074

13.086

0.000***

PG_AU_SEN_ASSOC

+

-0.222

-4.870

0.000***

-0.187

-1.663

0.097*

UG_MSA_AU_SEN_ASSOC

+

-0.156

-4.374

0.000***

-0.142

-1.560

0.119

UG_ELITE_AU_SEN_ASSOC

+

-0.016

-0.278

0.781

0.000

0.003

0.499

UG_AACSB_ACC_AU_SEN_ASSOC

+

0.206

4.359

0.000***

0.200

1.740

0.041**

AUDIT_CLIENTS

+

-0.169

-8.650

0.000***

-0.148

-3.072

0.002***

ICW

+

0.342

11.786

0.000***

0.240

3.314

0.000***

RESTATE

+

0.087

5.280

0.000***

0.078

1.727

0.042**

GCM

+

0.124

3.672

0.000***

0.242

2.813

0.003***

SIZE

+

0.463

128.681

0.000***

0.448

47.617

0.000***

LOSS

+

0.168

13.018

0.000***

0.130

3.985

0.000***

EQUITY_MULTIPLIER

+

0.000

0.866

0.194

0.001

0.899

0.185

QUICK

-

-0.025

-13.001

0.000***

-0.024

-4.413

0.000***

OPSEG

+

0.012

9.430

0.000***

0.032

4.199

0.000***

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

GEOSEG

+

0.008

11.125

0.000***

0.026

5.407

0.000***

lnPOPULATION

+

0.044

7.393

0.000***

0.041

2.662

0.004***

FIRM_TENURE

+

0.063

3.549

0.000***

-0.018

-0.425

0.671

NATIONAL_LEADER

+

-0.012

-1.234

0.217

0.007

0.285

0.388

CITY_LEADER

+

0.030

2.866

0.002***

0.034

1.261

0.104

NONAUDIT

+

0.017

12.435

0.000***

0.017

4.884

0.000***

FEM_AU

?

0.740

8.545

0.000***

0.564

2.585

0.010***

INITIAL

-

0.104

4.861

0.000***

0.081

1.280

0.201

FOREIGN

+

0.050

8.047

0.000***

0.019

1.472

0.071*

MERGER

+

0.055

5.230

0.000***

0.060

2.284

0.011**

DCFO

+

0.117

6.776

0.000***

0.124

2.696

0.004***

GROWTH

+

0.014

2.001

0.023**

0.017

0.779

0.218

Year and Industry fixed effects

Yes

Total Number of Observations

11,301

1,565

Model p-value

0.000***

R2

77.1%

78.8%

Expected Sign

Estimate

Regression of Audit Quality on the Education Quality of Audit Human Capital at the Audit Office Department (H1c)

140

When evaluating the results using only 2017-2018 control variable years, the findings

indicate a significant negative relationship between RESTATE and

UG_AACSB_ACC_AU_PART (p = 0.013) and significant positive relationships between

RESTATE and UG_ELITE_AU_SMGR_MGR (p < 0.01) and UG_MSA_AU_SEN_ASSOC

(p = 0.077), while none of the control variable results are unexpected.

Panels B1-B4 of Table 13 present the results from the logistic regressions of ICW

on the test and control variables for each staffing level (B1=partner, B2=principal,

B3=senior manager-manager, and B4=senior associate-associate). The results indicate

significant positive relationships between ICW and PG_AU_SMGR_MGR (p = 0.015)

and UG_MSA_AU_SMGR_MGR (p = 0.023). These results indicate that material

internal control weaknesses tend to be detected by audit departments with more senior

manager/manager audit human capital possessing post-graduate degrees and

undergraduate degrees from institutions located in the same metro-statistical area as their

home office. However, none of the educational test variables are significant at the

principal and partner levels, aligning with H1c in that the positive effects of education on

audit quality decrease as seniority levels increase. Of note is that the educational test

variables are not significant at the senior associate/associate level, possibly because

senior associates/associates tend to be involved in the audit detail testing, whereas senior

managers/managers tend to be more involved in the review process and in making overall

assessments of control deficiencies that would include assessing the presence of

compensating controls. The significant positive relationships between ICW and

AU_FEES, LOSS, and MERGER are expected. The significant negative relationships

between ICW and SIZE, QUICK, and INITIAL are also expected. When evaluating the

141

results using only 2017-2018 control variable years, the results show an unexpected

significant negative relationship between ICW and UG_MSA_AU_PRIN (p = 0.085),

while none of the control variable results are unexpected.

Panels C1-C4 of Table 13 present the results from the logistic regressions of

GCM on the test and control variables for each staffing level (C1=partner, C2=principal,

C3=senior manager-manager, and C4=senior associate-associate). The results indicate a

significant negative relationship between GCM and UG_AACSB_ACC_AU_PART (p =

0.063) and significant positive relationships between GCM and UG_ELITE_AU_PRIN (p

= 0.100), PG_AU_SMGR_MGR (p < 0.01), UG_AACSB_ACC_AU_SMGR_MGR (p =

0.078), and PG_AU_SEN_ASSOC (p = 0.042), partially aligning with H1c. The

significant negative relationship at the partner level is unexpected, although it may again

indicate that disproportionally staffed offices have inconsistent audit quality. The

significant positive relationships between GCM and AU_FEES, BANKRUPTCY, and

REPORTLAG are expected, as are the significant negative relationships between GCM

and SIZE, QUICK, MERGER, LnPOPULATION, GEOSEG, and NATIONAL_LEADER.

When evaluating the results using only 2017-2018 control variable years, the findings

indicate a significant negative relationship between GCM and UG_MSA_AU_PRIN,

while none of the control variable results are unexpected.

Panels D1-D4 of Table 13 present the results from the OLS regression of

AU_FEES on the test and control variables for each staffing level (D1=partner,

D2=principal, D3=senior manager-manager, and D4=senior associate-associate). The

results indicate significant negative relationships between AU_FEES and

UG_ELITE_AU_PRIN (p = 0.053), UG_AACSB_ACC_AU_PRIN (p = 0.048),

PG_AU_SMGR_MGR (p < 0.01), UG_MSA_AU_SMGR_MGR (p < 0.01),

142

UG_ELITE_AU_SMGR_MGR (p = 0.009), PG_AU_SEN_ASSOC (p < 0.01), and

UG_MSA_AU_SEN_ASSOC (p < 0.01), and significant positive relationships between

AU_FEES and PG_AU_PRIN (p < 0.01), UG_MSA_AU_PRIN (p = 0.021),

UG_AACSB_ACC_AU_SMGR_MGR (p = 0.032), and

UG_AACSB_ACC_AU_SEN_ASSOC (p < 0.01). The results seem to indicate that, while

educational effects on audit fees tend to disappear at the partner staffing level (as

expected by H1c), the educational effects of other staffing levels are mixed and should be

examined at a more detailed level in future research. The findings again report a

significant negative relationship between AU_FEES and AUDIT_CLIENTS (p < 0.01) at

all staffing levels, perhaps indicating that offices with greater amounts of audit clients per

auditor can create resource-sharing efficiencies among similar clients. The results also

indicate significant positive relationships between AU_FEES and RESTATE, ICW, GCM,

SIZE, LOSS, OP_SEG, GEO_SEG, LnPOPULATION, FIRM_TENURE, CITY_LEADER

(when controlling for NATIONAL_LEADER), NONAUDIT, INITIAL, FOREIGN,

MERGER, DCFO, and GROWTH, and a significant negative relationship between

AU_FEES and QUICK. Additionally, the results continue to indicate the presence of a

significant female audit fee premium when evaluating the positive relationship between

AU_FEES and FEM_AU (p < 0.01). When evaluating the results using only 2017-2018

control variable years, the results indicate significant negative relationships between

AU_FEES and PG_AU_SMGR_MGR (p = 0.001), UG_MSA_AU_SMGR_MGR (p =

0.001), UG_ELITE_AU_SMGR_MGR (p = 0.074), and PG_AU_SEN_ASSOC (p =

0.097), and significant positive relationships between AU_FEES and PG_AU_PRIN (p <

0.01) and UG_AACSB_ACC_AU_SEN_ASSOC (p = 0.041), while none of the control

variable results are unexpected.

143

Results Summary for Education Quality of Audit Human Capital at the Audit Office

Department and Audit Quality (H1c)

The test results for H1c first suggest that audit departments with higher

proportions of principals with post-graduate degrees; higher proportions of senior

managers/managers with elite undergraduate educations; and higher proportions of senior

associates/associates who attended an undergraduate program in the same metrostatistical

area as their current audit office all, surprisingly, have a significant positive relationship

to subsequent financial restatements, none of which align with H1c. The results also

indicate that offices with higher proportions of partners who attended undergraduate

accounting programs separately accredited by the AACSB have a significant negative

relationship to subsequent financial restatements, which does not fully align with H1c (as

H1c suggests that positive education effects are reduced at higher staffing levels). While

these mixed results warrant additional future research, a simple explanation might be that

audit departments that are staffed disproportionately (see Table 4 – the shape of a

descending slope) do not perform as well as audit departments that are staffed

proportionately. When considering only 2017-2018 control variable years, all of the test

variable results (except for PG_AU_PRIN which is no longer significant) are similar to

the results from the main model, while none of the control variable results are

unexpected.

The test results for H1c also suggest that material internal control weaknesses are

more likely to be reported by audit offices with higher proportions of senior

managers/managers who have a post-graduate degree or attended an undergraduate

144

program in the same metro-statistical area as their current audit office, which aligns with

H1c in that educational effects should lessen in correspondence with staffing level

increases, especially at the partner and principal levels. It could be that senior

managers/managers are involved more in the audit review processes where control

weaknesses would be identified, as opposed to associates who are more involved in detail

testing or partners and principals who might be more involved with the process of client

negotiations to report or withhold material internal control weakness disclosures. When

considering only 2017-2018 control variable years, only the model from Panel B2 is

significant, although none of the test variables are significant, while none of the control

variable results are unexpected.

The test results for H1c also suggest that audit departments with greater

proportions of partners who attended undergraduate accounting programs separately

accredited by the AACSB; greater proportions of principals who attended an elite

undergraduate institution; greater proportions of senior managers/managers with a

postgraduate degree or who attended undergraduate accounting programs separately

accredited by the AACSB; and greater proportions of senior associates/associates with a

post-graduate degree are all positively associated with going concern modifications.

These results partially align with H1c (warranting future research) by indicating that

general educational effects (i.e. the general possession of a post-graduate degree) are

more significant at the senior associate/associate and senior manager/manager levels,

whereas specialized educational effects (i.e. an elite education or a separately accredited

accounting program) may not be marginalized by the experience that comes with

advances in staffing level.

145

Finally, the test results for H1c indicate that audit fees do not have a significant

relationship with partner educational effects, which aligns with the expectations of H1c.

However, the test results also suggest that audit departments with greater proportions of

principals with post-graduate degrees or attended an undergraduate program in the same

metro-statistical area as their current audit office are associated with higher audit fees,

while audit departments with greater proportions of principals with elite educations or

who attended undergraduate accounting programs separately accredited by the AACSB

are associated with lower audit fees. In addition, audit departments with higher

proportions of both senior managers/managers and senior associates/associates who have

a post-graduate degree or attended an undergraduate program in the same metrostatistical

area as their current audit office are associated with lower audit fees, while audit

departments with higher proportions of both senior managers/managers and senior

associates/associates who attended undergraduate accounting programs separately

accredited by the AACSB are associated with higher audit fees. Taken together, these

mixed results might be capturing inefficiency effects that exists in offices which are

disproportionately staffed, or capturing the effects of offices that borrow (lend) audit

human capital from (to) other offices with excess (limited) capacity. Future research in

this area could attempt to uncover which offices tend to lend audit human capital and

which offices tend to borrow audit human capital, and how educational background

affects this practice. When evaluating the test results for H1c using only control variables

from 2017-2018, the results for senior managers/managers and senior

associates/associates are similar to the main model, while only the proportion of

146

principals with post-graduate degrees remains significant, possibly indicating that for the

control sample, educational effects are stronger for lower staffing levels. None of the

2017-2018 control variable results are unexpected.

Logistic and OLS Regressions – Audit Personnel Experience

Auditor Experience at the Audit Office Department and Audit Quality (H2a)

Table 14 provides the regression results of the relationship between audit

personnel experience at the audit office department and audit quality for Hypotheses 2a.

Panel A of Table 14 presents the results from the logistic regression of RESTATE on the

test and control variables. The results indicate a significant negative association between

RESTATE and SMGR_MGR_EXP_DEPT_AVG (p = 0.016). This result could indicate

that experienced managers have an impact on audit quality (measured by subsequent

restatements), while experienced staff or experienced principals and partners may not.

Although this finding does not fully align with H2a, the assessment is intuitive, as

managers are the primary database and financial statement reviewers in an audit

engagement, and therefore might have the most influence on the incorrect application of

a standard that would require a subsequent restatement. Additionally, the results continue

to indicate a significant negative association between RESTATE and FEM_AU (0.026).

The results also indicate significant positive relationships between RESTATE and ICW,

AU_FEES, EQUITY_MULTIPLIER, and FIRM_TENURE, and significant negative

relationships between RESTATE and AUDIT_CLIENTS, NATIONAL_LEADER, and

147

SIZE, all of which are expected based on the prior literature. When evaluating the results

using only 2017-2018 control variable years, the test variables are not significant, while

none of the control variable results are unexpected.

Panel B of Table 14 presents the results from the logistic regression of ICW on the

test and control variables for Hypothesis 2a. The results indicate significant positive

relationships between ICW and PART_EXP_DEPT_AVG (p = 0.021) and

SEN_ASSOC_EXP_DEPT_AVG (p = 0.087), and a significant negative relationship

between ICW and PRIN_EXP_DEPT_AVG (p = 0.014), which partially aligns with H2a,

in that experience seems to impacts audit quality much more at senior levels. However,

the negative coefficient for principals is unexpected when compared to the positive

coefficient for partners. It could be that some of the audit principals act in an

administrative audit or director role, such as a revenue auditing expert or PCAOB

compliance expert, and give general guidance rather than work on a specific audit

engagement at an office. When comparing to the results from Panel B1 of Table 13

(H1c), it is interesting to note that the coefficients for audit department partners/education

interactions are negative, while the coefficient for the experience of the audit department

partners (PART_EXP_DEPT_AVG) is positive, possibly indicating that the experience

level of partners (rather than simply headcount) has a more powerful effect in the

reporting of material internal control weaknesses.

TABLE 14 - Regression of Audit Quality on Auditor Experience at the Audit Office Department (H2a)

Panel A: Using RESTATE as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Variable Wald

Intercept

?

-2.038

2.981

0.084*

-21.531

0.000

0.999

PART_EXP_DEPT_AVG

-

-0.004

0.374

0.271

-0.018

0.628

0.214

PRIN_EXP_DEPT_AVG

-

0.012

2.590

0.108

-0.027

1.268

0.130

SMGR_MGR_EXP_DEPT_AVG

-

-0.015

4.646

0.016**

-0.036

1.468

0.113

SEN_ASSOC_EXP_DEPT_AVG

-

-0.044

0.120

0.365

-0.414

0.879

0.174

AUDIT_CLIENTS

-

-0.358

4.984

0.013**

-0.141

0.110

0.370

ICW

+

1.413

107.827

0.000***

2.190

31.403

0.000***

AU_FEES

+

0.338

24.558

0.000***

0.296

1.907

0.084*

SIZE

-

-0.172

18.341

0.000***

-0.220

2.977

0.042**

LOSS

+

0.005

0.003

0.479

0.469

2.863

0.091*

EQUITY_MULTIPLIER

+

0.007

4.075

0.022**

0.006

0.237

0.313

QUICK

-

-0.023

2.104

0.074*

-0.047

0.675

0.206

OPSEG

+

0.006

0.539

0.232

0.023

0.149

0.350

GEOSEG

+

0.001

0.015

0.451

0.001

0.489

Expected Sign

Estimate

Wald

p

-

value

Estimate

p

-

value

Regression of Audit Quality on Auditor Experience at the Audit Office Department (H2a)

LnPOPULATION

-

-0.010

0.059

0.405

0.172

1.792

0.181

FIRM_TENURE

+

0.240

3.958

0.024**

0.153

0.170

0.340

NATIONAL_LEADER

-

-0.163

4.887

0.014**

-0.162

0.532

0.233

CITY_LEADER

-

0.000

0.996

0.161

0.491

0.484

NONAUDIT

+

0.009

0.763

0.191

0.034

1.026

0.156

FEM_AU

?

-1.370

4.928

0.026**

-0.955

0.254

0.614

INITIAL

+

0.000

0.500

0.732

2.250

0.067*

MERGER

+

-0.008

0.013

0.910

0.100

0.208

0.324

CHANGE_IN_RECEIVABLES

+

-0.001

0.121

0.728

-0.021

0.022

0.881

CHANGE_IN_INVENTORIES

+

-0.005

0.029

0.865

-0.004

0.006

0.937

CHANGE_IN_CASH_SALES

+

-0.002

0.986

0.321

-0.071

1.721

0.190

Year and Industry fixed effects

Yes

Number of Restatements

1,025

122

Total Number of Observations

11,301

1,524

Model p-value

0.000***

Pseudo R2

5.8%

19.9%

Panel B: Using ICW as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Regression of Audit Quality on Auditor Experience at the Audit Office Department (H2a)

Variable Wald

Intercept

?

-26.239

0.000

0.998

-19.325

0.000

0.999

PART_EXP_DEPT_AVG

+

0.025

4.180

0.021**

0.062

3.559

0.030**

PRIN_EXP_DEPT_AVG

+

-0.036

6.045

0.014**

-0.006

0.028

0.867

SMGR_MGR_EXP_DEPT_AVG

+

-0.020

2.453

0.117

-0.037

1.679

0.195

SEN_ASSOC_EXP_DEPT_AVG

+

0.282

1.856

0.087*

-0.562

0.690

0.406

AUDIT_CLIENTS

+

0.073

0.120

0.365

-1.163

1.406

0.236

AU_FEES

+

1.443

143.096

0.000***

1.38

13.328

0.000***

SIZE

-

-0.798

125.090

0.000***

-0.856

15.323

0.000***

LOSS

+

0.201

2.005

0.079*

0.123

0.104

0.374

EQUITY_MULTIPLIER

+

0.006

0.962

0.164

-0.009

0.283

0.595

QUICK

-

-0.042

1.548

0.107

-0.215

1.985

0.080*

OPSEG

+

-0.012

0.526

0.468

-0.007

0.002

0.961

GEOSEG

+

0.002

0.030

0.432

-0.039

0.283

0.595

LnPOPULATION

-

0.014

0.036

0.850

-0.275

1.678

0.098*

FIRM_TENURE

-

0.244

1.458

0.227

0.599

1.325

0.250

NATIONAL_LEADER

+

-0.078

0.350

0.554

-0.023

0.004

0.947

Expected Sign

Estimate

Wald

p

-

value

Estimate

p

-

value

Regression of Audit Quality on Auditor Experience at the Audit Office Department (H2a)

CITY_LEADER

+

-0.178

1.687

0.194

0.138

0.143

0.353

NONAUDIT

+

0.006

0.094

0.380

-0.002

0.002

0.961

FEM_AU

?

1.477

1.813

0.178

4.231

1.610

0.204

INITIAL

-

-0.325

1.638

0.101

0.142

0.043

0.836

MERGER

+

0.273

4.556

0.017**

0.609

3.030

0.041**

Year and Industry fixed effects

Yes

Number of Control Weaknesses

308

45

Total Number of Observations

11,301

1,565

Model p-value

0.000***

0.074*

Pseudo R2

13.8%

24.9%

Panel C: Using GCM as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Variable Wald p-value Wald

Intercept

?

10.699

14.553

0.000***

8.714

0.000

1.000

PART_EXP_DEPT_AVG

+

0.006

0.093

0.381

-0.062

0.625

0.429

Expected Sign

Estimate

p

-

value

Regression of Audit Quality on Auditor Experience at the Audit Office Department (H2a)

PRIN_EXP_DEPT_AVG

+

-0.002

0.012

0.911

-0.067

0.870

0.351

SMGR_MGR_EXP_DEPT_AVG

+

-0.058

4.730

0.030**

-0.283

1.351

0.245

SEN_ASSOC_EXP_DEPT_AVG

+

-0.051

0.028

0.866

-0.246

0.065

0.799

AUDIT_CLIENTS

-

-0.416

1.211

0.136

1.372

1.225

0.268

ICW

+

0.069

0.023

0.440

-15.544

0.000

0.998

AU_FEES

+

0.188

1.909

0.084*

0.061

0.008

0.464

SIZE

-

-0.477

25.661

0.000***

-0.649

2.016

0.078*

LOSS

+

0.384

0.537

0.232

18.661

0.000

0.499

EQUITY_MULTIPLIER

+

-0.004

0.187

0.665

-0.033

0.270

0.603

QUICK

-

-0.166

19.000

0.000***

-0.092

0.782

0.189

OPSEG

-

0.013

0.122

0.727

-0.083

0.053

0.410

GEOSEG

-

-0.057

6.379

0.006***

0.189

1.955

0.162

LnPOPULATION

-

-0.200

2.994

0.042**

-0.941

4.281

0.020**

FIRM_TENURE

-

0.329

1.986

0.159

0.713

0.734

0.392

NATIONAL_LEADER

+

-0.320

2.838

0.092*

-0.843

1.475

0.225

CITY_LEADER

+

0.098

0.270

0.302

0.054

0.006

0.470

NONAUDIT

+

0.001

0.005

0.473

-0.042

0.539

0.463

Regression of Audit Quality on Auditor Experience at the Audit Office Department (H2a)

FEM_AU

?

2.696

3.262

0.071*

-1.712

0.123

0.726

INITIAL

-

-0.217

0.702

0.201

-0.641

0.325

0.285

MERGER

-

-0.407

2.357

0.063*

-0.891

0.741

0.195

REPORTLAG

+

0.002

1.613

0.102

0.065

5.155

0.012**

BANKRUPTCY

-

-0.067

32.554

0.000***

-0.085

3.947

0.024**

Year and Industry fixed effects

Yes

Number of Going Concern Modifications

235

19

Total Number of Observations

3,705

481

Model p-value

0.000***

Pseudo R2

40.5%

60.0%

Panel D: Using AU_FEES as the Dependent Variable

Control Variable Years: 2011-2018

Control Variable Years: 2017-2018

Variable

t-value

p-value

t-value

p-value

Intercept

?

3.539

28.531

0.000***

4.146

13.065

0.000***

PART_EXP_DEPT_AVG

+

-0.004

-3.948

0.000***

-0.004

-1.560

0.119

PRIN_EXP_DEPT_AVG

+

0.000

-0.376

0.354

-0.002

-0.913

0.361

Regression of Audit Quality on Auditor Experience at the Audit Office Department (H2a)

SMGR_MGR_EXP_DEPT_AVG

+

0.003

4.135

0.000***

0.006

2.917

0.002***

SEN_ASSOC_EXP_DEPT_AVG

+

-0.012

-0.649

0.517

-0.092

-2.043

0.041**

AUDIT_CLIENTS

+

-0.161

-8.378

0.000***

-0.129

-2.764

0.006***

ICW

+

0.345

11.850

0.000***

0.249

3.443

0.000***

RESTATE

+

0.086

5.217

0.000***

0.072

1.583

0.057*

GCM

+

0.125

3.721

0.000***

0.242

2.827

0.003***

SIZE

+

0.464

128.695

0.000***

0.448

47.808

0.000***

LOSS

+

0.169

13.074

0.000***

0.129

3.972

0.000***

EQUITY_MULTIPLIER

+

0.001

0.883

0.189

0.001

0.965

0.168

QUICK

-

-0.026

-13.367

0.000***

-0.024

-4.528

0.000***

OPSEG

+

0.011

9.126

0.000***

0.031

4.151

0.000***

GEOSEG

+

0.008

11.106

0.000***

0.026

5.375

0.000***

lnPOPULATION

+

0.045

7.636

0.000***

0.047

3.132

0.001***

FIRM_TENURE

+

0.060

3.409

0.000***

-0.024

-0.550

0.583

NATIONAL_LEADER

+

-0.017

-1.710

0.087*

-0.003

-0.103

0.918

CITY_LEADER

+

0.030

2.847

0.002***

0.034

1.280

0.101

NONAUDIT

+

0.017

12.525

0.000***

0.017

4.886

0.000***

Regression of Audit Quality on Auditor Experience at the Audit Office Department (H2a)

FEM_AU

?

0.623

7.201

0.000***

0.499

2.295

0.022**

INITIAL

-

0.103

4.821

0.000***

0.088

1.394

0.164

FOREIGN

+

0.050

7.948

0.000***

0.016

1.260

0.104

MERGER

+

0.052

4.939

0.000***

0.061

2.340

0.001***

DCFO

+

0.114

6.602

0.000***

0.115

2.508

0.006***

GROWTH

+

0.014

2.009

0.023**

0.020

0.933

0.176

Year and Industry fixed effects

Yes

Total Number of Observations

11,301

1,565

Model p-value

0.000***

R2

77.1%

78.9%

Expected Sign

Estimate

Regression of Audit Quality on Auditor Experience at the Audit Office Department (H2a)

157

The significant positive relationships between ICW and AU_FEES, MERGER, and LOSS

are expected, as is the significant negative relationship between ICW and SIZE. When

evaluating the results using only 2017-2018 control variable years, the significant

relationship between ICW and PART_EXP_DEPT_AVG remains, while none of the

control variable results are unexpected.

Panel C of Table 14 presents the results from the logistic regression of GCM on

the test and control variables for Hypothesis 2a. The results show a significant negative

relationship between GCM and SMGR_MGR_EXP_DEPT_AVG (p = 0.030). At first

glance this result is unexpected. However, the decision of whether or not to modify an

audit opinion for a going concern risk is often based on complex calculations performed

during the manager review stage, and pressures caused by deadlines and workload

compressions could have a greater effect on manager audit quality than simply years of

experience. The significant negative relationships between GCM and SIZE, QUICK,

GEOSEG, NATIONAL_LEADER, LnPOPULATION, and BANKRUPTCY are expected,

as is the significant positive relationship between GCM and AU_FEES. Of note is the

significant positive relationship between GCM and FEM_AU (p = 0.071), indicating that

offices with greater proportions of female auditors may be more likely to report going

concern opinion modifications than offices with greater proportions of male auditors.

When evaluating the results using only 2017-2018 control variable years, none of the test

variables are significant, while none of the control variable results are unexpected.

Panel D of Table 14 presents the results from the OLS regression of AU_FEES on

the test and control variables for Hypotheses 2a. The results show a significant negative

158

relationship between AU_FEES and PART_EXP_DEPT_AVG (p < 0.01), and a

significant positive relationship between AU_FEES and SMGR_MGR_EXP_DEPT_AVG

(p < 0.01). These mixed results are unexpected and, similar to Panel D3 of Table 13,

could indicate that senior managers and managers are more involved in the audit

processes and that the Big 4 bill excess fees for their expertise, or alternatively that

disproportionately staffed offices are inefficient. The findings indicate significant

positive relationships between AU_FEES and RESTATE, ICW, SIZE, LOSS, OP_SEG,

GEO_SEG, LnPOPULATION, FIRM_TENURE, CITY_LEADER (when controlling for

NATIONAL_LEADER), NONAUDIT, INITIAL, FOREIGN, MERGER, DCFO, and

GROWTH, and a significant negative relationship between AU_FEES and QUICK. The

findings again report a significant negative relationship between AU_FEES and

AUDIT_CLIENTS (p < 0.01) at all staffing levels, similar to previously reported results,

as well as the presence of a significant female audit fee premium when evaluating the

positive relationship between AU_FEES and FEM_AU (p < 0.01). When evaluating the

results using only 2017-2018 control variable years, the findings continue to show a

significant positive relationship between AU_FEES and SMGR_MGR_EXP_DEPT_AVG,

while also showing a significant negative relationship with

SEN_ASSOC_EXP_DEPT_AVG, which is unexpected but could indicate that experience

is not as valuable at the senior associate and associate levels, or perhaps that more

experienced senior associates and associates are being shared within the office or among

other offices in order to create engagement efficiencies. Additionally, none of the control

variable results are unexpected.

159

Auditor Experience at the Audit Firm and Audit Quality (H2a)

Table 15 provides the regression results of the relationship between audit

personnel experience at the firm level and audit quality for Hypotheses 2a. Panel A of

Table 15 presents the results from the logistic regression of RESTATE on the test and

control variables. Similar to Panel A of Table 14, the results continue to indicate a

significant negative association between RESTATE and SMGR_MGR_EXP_FIRM_AVG

(p = 0.039). Unlike Panel A of Table 14, however, the results show a significant positive

association between RESTATE and PRIN_EXP_FIRM_AVG (p = 0.069), which is

unexpected and therefore warrants future research. It is possible that principals in the

national office or in expert offices are utilized on audit engagements that require

restatements, or that, as previously mentioned, disproportionately staffed offices lead to

poor quality because of inefficiencies and workload volume. The results continue to

indicate a significant negative association between RESTATE and FEM_AU. The results

also indicate significant positive relationships between RESTATE and ICW, AU_FEES,

FIRM_TENURE, and EQUITY_MULTIPLIER and significant negative relationships

between RESTATE and AUDIT_CLIENTS, SIZE, QUICK, and NATIONAL_LEADER, all

of which are expected. When evaluating the results using only 2017-2018 control

variable years, none of the test variables are significant, while none of the control

variable results are unexpected.

160

Panel B of Table 15 presents the results from the logistic regression of ICW on the

test and control variables for Hypothesis 2a. As expected, the results indicate a

significant positive relationship between ICW and PART_EXP_FIRM_AVG (p = 0.046).

However, similar to Panel A of Table 15, the significant negative relationship between

ICW and PRIN_EXP_FIRM_AVG (p = 0.060) is unexpected and warrants future research.

The significant positive relationships between ICW and AU_FEES, MERGER, and LOSS

are expected, as are the significant negative relationships between ICW and SIZE,

QUICK, and INITIAL. When evaluating the results using only 2017-2018 control

variable years, the significant negative relationship between ICW and

PART_EXP_DEPT_AVG remains, while none of the control variable results are

unexpected.

Panel C of Table 15 presents the results from the logistic regression of GCM on

the test and control variables for Hypothesis 2a. The results show an unexpected

significant negative relationship between GCM and SMGR_MGR_EXP_FIRM_AVG (p =

0.041), similar to Panel C of Table 14. The significant negative relationships between

GCM and BANKRUPTCY, GCM, SIZE, QUICK, GEOSEG, LnPOPULATION, and

NATIONAL_LEADER are all expected, as is the significant positive relationship between

GCM and AU_FEES. When evaluating the results using only 2017-2018 control variable

years, the test variables are not significant, while none of the control variable results are

unexpected.

Panel D of Table 15 presents the results from the OLS regression of AU_FEES on

the test and control variables for Hypotheses 2a.2. The results show a significant

161

negative relationship between AU_FEES and PART_EXP_FIRM_AVG (p < 0.01), and

significant positive relationships between AU_FEES and PRIN_EXP_FIRM_AVG (p =

0.002), SMGR_MGR_EXP_FIRM_AVG (p < 0.01), and SEN_ASSOC_EXP_FIRM_AVG

(p = 0.002).

TABLE 15 - Regression of Audit Quality on Auditor Experience at the Audit Office Firm (H2a)

Panel A: Using RESTATE as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Variable p-value Wald

Intercept

?

-2.311

4.024

0.045**

-22.298

0.000

0.999

PART_EXP_FIRM_AVG

-

-0.005

0.472

0.246

-0.019

0.720

0.198

PRIN_EXP_FIRM_AVG

-

0.010

3.304

0.069*

-0.005

0.081

0.388

SMGR_MGR_EXP_FIRM_AVG

-

-0.010

3.112

0.039**

-0.021

0.966

0.163

SEN_ASSOC_EXP_FIRM_AVG

-

-0.007

0.007

0.468

-0.091

0.133

0.358

AUDIT_CLIENTS

-

-0.349

4.753

0.015**

-0.117

0.076

0.392

ICW

+

1.413

107.916

0.000***

2.169

30.950

0.000***

AU_FEES

+

0.332

23.729

0.000***

0.297

1.917

0.083*

SIZE

-

-0.170

17.798

0.000***

-0.215

2.843

0.046**

LOSS

+

0.006

0.004

0.474

0.485

3.089

0.079*

EQUITY_MULTIPLIER

+

0.007

3.974

0.023**

0.005

0.160

0.345

QUICK

-

-0.023

2.165

0.071*

-0.045

0.629

0.214

OPSEG

+

0.006

0.501

0.240

0.027

0.197

0.329

GEOSEG

+

0.001

0.023

0.441

0.002

0.003

0.478

Expected Sign

Estimate

Wald

Estimate

p

-

value

(H2a)

LnPOPULATION

-

0.001

0.000

0.987

0.169

1.721

0.190

FIRM_TENURE

+

0.237

3.877

0.025**

0.156

0.175

0.338

NATIONAL_LEADER

-

-0.156

4.493

0.017**

-0.135

0.375

0.270

CITY_LEADER

-

-0.002

0.000

0.492

0.183

0.635

0.426

NONAUDIT

+

0.009

0.729

0.197

0.034

1.037

0.154

FEM_AU

?

-1.275

4.218

0.040**

-1.441

0.579

0.447

INITIAL

+

0.003

0.000

0.493

0.705

2.082

0.075*

MERGER

+

-0.009

0.016

0.900

0.086

0.153

0.348

CHANGE_IN_RECEIVABLES

+

-0.001

0.126

0.723

-0.025

0.028

0.868

CHANGE_IN_INVENTORIES

+

-0.005

0.028

0.867

-0.006

0.011

0.917

CHANGE_IN_CASH_SALES

+

-0.002

1.000

0.317

-0.070

1.727

0.189

Year and Industry fixed effects

Yes

Number of Restatements

1,025

122

Total Number of Observations

11,301

1,524

Model p-value

0.000***

Pseudo R2

5.7%

19.3%

Regression of Audit Quality on Auditor Experience at the Audit Office Firm

(H2a)

Panel B: Using ICW as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Variable Wald

Intercept

?

-25.601

0.000

0.998

-19.728

0.000

0.999

PART_EXP_FIRM_AVG

+

0.021

2.849

0.046**

0.053

2.511

0.057*

PRIN_EXP_FIRM_AVG

+

-0.023

3.533

0.060*

-0.036

1.243

0.265

SMGR_MGR_EXP_FIRM_AVG

+

-0.013

1.632

0.201

-0.025

1.172

0.279

SEN_ASSOC_EXP_FIRM_AVG

+

-0.023

0.027

0.871

-0.412

0.999

0.317

AUDIT_CLIENTS

+

0.135

0.429

0.257

-1.18

1.427

0.232

AU_FEES

+

1.452

143.866

0.000***

1.381

13.466

0.000***

SIZE

-

-0.802

126.798

0.000***

-0.849

15.159

0.000***

LOSS

+

0.192

1.821

0.089*

0.14

0.134

0.357

EQUITY_MULTIPLIER

+

0.006

0.924

0.168

-0.01

0.392

0.531

QUICK

-

-0.045

1.747

0.093*

-0.206

1.889

0.085*

OPSEG

+

-0.012

0.509

0.475

-0.015

0.011

0.915

GEOSEG

+

0.001

0.017

0.449

-0.032

0.190

0.663

LnPOPULATION

-

0.010

0.019

0.890

-0.25

1.427

0.116

FIRM_TENURE

-

0.241

1.415

0.234

0.616

1.405

0.236

NATIONAL_LEADER

+

-0.112

0.738

0.390

-0.007

0.000

0.984

Expected Sign

Estimate

Wald

p

-

value

Estimate

p

-

value

(H2a)

CITY_LEADER

+

-0.153

1.254

0.263

0.09

0.061

0.403

NONAUDIT

+

0.005

0.075

0.393

0.001

0.491

FEM_AU

?

1.673

2.238

0.135

3.899

1.425

0.233

INITIAL

-

-0.340

1.784

0.091*

0.131

0.037

0.848

MERGER

+

0.271

4.477

0.017**

0.57

2.670

0.051*

Year and Industry fixed effects

Yes

Number of Control Weaknesses

308

45

Total Number of Observations

11,301

1,565

Model p-value

0.000***

0.072*

Pseudo R2

13.5%

25.0%

Regression of Audit Quality on Auditor Experience at the Audit Office Firm

Panel C: Using GCM as the Dependent Variable Control Variable Years: 2011-2018 Control Variable Years: 2017-2018

Variable Wald p-value Wald

Intercept

?

10.341

14.044

0.000***

9.629

0.000

1.000

Expected Sign

Estimate

p

-

value

(H2a)

PART_EXP_FIRM_AVG

+

-0.001

0.005

0.945

-0.099

1.036

0.309

PRIN_EXP_FIRM_AVG

+

-0.004

0.042

0.837

0.031

0.318

0.287

SMGR_MGR_EXP_FIRM_AVG

+

-0.042

4.187

0.041**

-0.278

0.955

0.328

SEN_ASSOC_EXP_FIRM_AVG

+

0.235

1.309

0.127

0.631

0.669

0.207

AUDIT_CLIENTS

-

-0.411

1.298

0.127

1.573

1.359

0.244

ICW

+

0.061

0.017

0.448

-16.056

0.000

0.998

AU_FEES

+

0.186

1.865

0.086*

0.154

0.054

0.409

SIZE

-

-0.481

26.016

0.000***

-0.770

3.088

0.040**

LOSS

+

0.429

0.671

0.207

19.268

0.000

0.499

EQUITY_MULTIPLIER

+

-0.004

0.191

0.662

-0.019

0.083

0.773

QUICK

-

-0.165

18.864

0.000***

-0.084

0.654

0.210

OPSEG

-

0.011

0.082

0.774

-0.188

0.255

0.307

GEOSEG

-

-0.055

6.086

0.007***

0.149

1.336

0.248

LnPOPULATION

-

-0.204

3.232

0.036**

-1.016

4.932

0.013**

FIRM_TENURE

-

0.336

2.084

0.149

1.012

1.545

0.214

NATIONAL_LEADER

+

-0.323

2.919

0.088*

-0.801

1.330

0.249

CITY_LEADER

+

0.132

0.487

0.243

0.125

0.030

0.431

(H2a)

NONAUDIT

+

0.001

0.003

0.480

-0.036

0.376

0.540

FEM_AU

?

2.200

2.361

0.124

-1.993

0.160

0.689

INITIAL

-

-0.220

0.726

0.197

-0.903

0.640

0.212

MERGER

-

-0.414

2.446

0.059*

-0.833

0.619

0.216

REPORTLAG

+

0.002

1.493

0.111

0.051

3.305

0.035**

BANKRUPTCY

-

-0.067

32.470

0.000***

-0.084

3.915

0.024**

Year and Industry fixed effects

Yes

Number of Going Concern Modifications

235

19

Total Number of Observations

3,705

481

Model p-value

0.000***

Pseudo R2

40.4%

59.9%

Regression of Audit Quality on Auditor Experience at the Audit Office Firm

Panel D: Using AU_FEES as the Dependent Variable

Control Variable Years: 2011-2018

Control Variable Years: 2017-2018

Variable

t-value

p-value

t-value

p-value

Intercept

?

3.454

28.853

0.000***

3.962

13.003

0.000***

(H2a)

PART_EXP_FIRM_AVG

+

-0.004

-3.935

0.000***

-0.005

-2.047

0.041**

PRIN_EXP_FIRM_AVG

+

0.002

2.892

0.002***

0.001

0.645

0.260

SMGR_MGR_EXP_FIRM_AVG

+

0.003

3.692

0.000***

0.004

2.520

0.006***

SEN_ASSOC_EXP_FIRM_AVG

+

0.033

2.891

0.002***

0.001

0.033

0.487

AUDIT_CLIENTS

+

-0.161

-8.356

0.000***

-0.126

-2.676

0.008***

ICW

+

0.345

11.885

0.000***

0.250

3.458

0.000***

RESTATE

+

0.084

5.141

0.000***

0.076

1.676

0.047**

GCM

+

0.125

3.719

0.000***

0.242

2.817

0.003***

SIZE

+

0.464

129.012

0.000***

0.449

47.941

0.000***

LOSS

+

0.169

13.093

0.000***

0.127

3.908

0.000***

EQUITY_MULTIPLIER

+

0.000

0.825

0.205

0.001

0.867

0.193

QUICK

-

-0.025

-13.192

0.000***

-0.024

-4.535

0.000***

OPSEG

+

0.011

9.268

0.000***

0.032

4.227

0.000***

GEOSEG

+

0.008

11.035

0.000***

0.026

5.336

0.000***

lnPOPULATION

+

0.042

7.170

0.000***

0.044

2.966

0.002***

FIRM_TENURE

+

0.059

3.356

0.000***

-0.025

-0.572

0.567

NATIONAL_LEADER

+

-0.012

-1.208

0.227

0.007

0.264

0.396

(H2a)

CITY_LEADER

+

0.032

2.989

0.002***

0.037

1.363

0.087*

NONAUDIT

+

0.017

12.357

0.000***

0.016

4.783

0.000***

FEM_AU

?

0.597

6.860

0.000***

0.440

2.007

0.045**

INITIAL

-

0.104

4.856

0.000***

0.084

1.325

0.186

FOREIGN

+

0.050

7.904

0.000***

0.017

1.283

0.100*

MERGER

+

0.052

4.982

0.000***

0.058

2.220

0.014**

DCFO

+

0.114

6.615

0.000***

0.123

2.671

0.004***

GROWTH

+

0.014

1.998

0.023**

0.018

0.853

0.197

Year and Industry fixed effects

Yes

Total Number of Observations

11,301

1,565

Model p-value

0.000***

R2

77.1%

78.8%

Expected Sign

Estimate

125

These mixed results capture the effects of inter-office firm experience and might indicate

that principals, senior managers/managers, and senior associates/associates are more

involved in the audit processes and that the Big 4 bill excess fees for their expertise at the

firm. The results also continue to suggest that future research regarding staffing

proportions and audit efficiency should be conducted. The findings indicate significant