The Effects of PCAOB Inspections on Auditor-Client Relationships

Andrew A. Acito

Eli Broad College of Business, Michigan State University

Chris E. Hogan

Eli Broad College of Business, Michigan State University

Richard D. Mergenthaler

Henry B. Tippie College of Business, The University of Iowa

Preliminary draft. Please do not circulate or cite without permission.

Revised: September 2013

Abstract

We investigate the effects of PCAOB inspections on the relationships between Big 4 auditors and their

clients using a new measure to capture the information garnered from the reports. Specifically, we

measure the relative importance of accounting standards to each client’s financial statements and

calculate the exposure to deficient auditing by relating their auditor’s inspection deficiencies to the

accounting standards. This measure of deficient auditing exposure is then adjusted by the exposure that

would occur for the average of the other Big 4 auditors. We find that our measure of relative exposure to

deficient auditing is positively related to auditor changes, but is not related to audit changes in audit fees.

These results suggest PCAOB inspections affect auditor-client relationships, but auditors do not have the

ability to increase fees to remediate deficient auditing, nor do they reduce fees to retain clients when they

have more deficiencies in areas important to the clients. Our findings have implications for understanding

and regulating the market for audit services.

We thank Jay Newquist and Kyle Peterson for their programming assistance and Courtney Shemka and June Sun for

their help with data collection. We also thank Brown Bag Participants at Michigan State University.

1

The Effects of PCAOB Inspections on Auditor-Client Relationships

1. Introduction

One of the most important aspects of the Sarbanes-Oxley Act of 2002 has been the creation of the

Public Company Accounting Oversight Board (PCAOB). The PCAOB is charged with establishing

standards for auditing public companies, inspecting auditors with publically held clients, and levying

penalties on auditors who do not comply with PCAOB standards. In 2005 testimony to the U.S. House of

Representatives, former PCAOB Chairman James McDonough emphasized the importance of auditor

inspections stating, “…the more significant, long-term effects of our work will be the product of our

oversight activities” (PCAOB 2005). Auditors, however, have sometimes been critical of what the

PCAOB identifies as deficiencies in its inspection reports. Deloitte, for example, responded to its 2007

PCAOB inspection report by saying, “reasonable judgments should be respected and not second-guessed"

(Johnson 2009). While the PCAOB’s inspection process has clearly led to significant changes in the way

audits are conducted, there has been little research that investigates the effects of reporting inspection

deficiencies.

We investigate whether PCAOB inspection reports affect auditor-client relationships. First, we

look for evidence of audit fee changes related to inspection issues. When the PCAOB identifies that an

audit is deficient in a particular area, the auditor is expected to take action to address the issue. The

auditor may work to improve auditing in the area of the deficiency, both at the client where the issue was

identified and at other clients, and this additional work may lead to higher fees. For example, the 2006

inspection report of PricewaterhouseCoopers states that at Issuer A, “The Firm failed to test the fair value

of warrants and stock-based compensation issued in two significant transactions during the year.”1 To

improve in this area, PricewaterhouseCoopers may revise its audit procedures not only for Issuer A, but

also for other issuers or clients that have high levels of warrants and stock-based transactions. Auditors

1 The PCAOB does not disclose the names of the clients where audit deficiencies are identified.

2

have a strong incentive to address their deficiencies because the PCAOB may see a failure to do so as a

lack of quality control. The PCAOB also evaluates auditors’ quality control systems, and can make these

deficiencies public if they are not resolved within one year (PCAOB 2012).2 Another possibility is that

an auditor will see the PCAOB’s attention on the area as an indication of higher audit risk and thus raise

fees on clients that have exposure to the area where the deficiency was identified. Conversely, audit fees

may decrease if auditors reduce fees in an attempt to retain the clients. An Auditor may agree to lower

fees to pacify a client that might otherwise dismiss the auditor because of a negative PCAOB inspection

report related to the client’s accounting (Abbot, Gunny, and Zhang 2013).

Second, we investigate whether PCAOB inspection reports affect the likelihood of Big 4 auditor

changes. A client may choose to dismiss an auditor that receives a negative PCAOB inspection report

(Abbot, Gunny, and Zhang 2013) and may be especially likely to dismiss their auditor or not engage a

new auditor if the auditor has deficiencies in an area that is important to the client’s accounting. The

PCAOB specifically suggests that audit committees ask auditors about deficiencies identified in their

audit as well as in audits of other clients that have similar accounting to the client (PCAOB 2012). It is

also possible that auditors want to avoid or discontinue relationships with clients that have high exposure

to certain accounting standards. Auditors may also be more likely to resign from or refuse to take on

marginal clients that use accounting standards that increase the auditors’ exposure to areas where

inspection deficiencies are likely.

An alternative to the hypothesized relations discussed above is that PCAOB inspection reports do

not affect auditor-client relationships for Big 4 auditors. There are reasons to believe that the PCAOB

reports are not informative to clients. For example, Lennox and Pittman (2010) find no evidence of a

relation between the number of issues identified in PCAOB inspections and changes in auditor market

2 As an example, in the 2008 and 2009 inspection reports for PricewaterhouseCoopers (PwC), the PCAOB

inspectors noted deficiencies related to auditing fair value measurements. Subsequently, Part II of those inspection

reports were released because the PCAOB felt PwC had not made sufficient progress in addressing quality controls

with respect to auditing fair value measurements, despite PwC taking actions which “included providing our audit

professionals with enhanced audit tools, training and additional technical guidance to promote more consistent audit

execution.” Of the Big 4 audit firms, only KPMG has yet to have a Part II of an annual inspection report released.

3

share. Gunny and Zhang (2013) also suggest inspection findings do not inform about audit quality for

larger audit firms, which the PCAOB inspects annually. Further, the PCAOB reports that auditors often

dismiss inspection findings in their communications with audit committees (PCAOB 2012).

To investigate the relation between PCAOB inspection results and auditor-client relationships, we

examine PCAOB inspection reports issued between 2005 and 2011 for Big 4 auditors. We search each

inspection report for a list of keywords developed by Folsom, Hribar, Mergenthaler, and Peterson (2012)

that relate to each accounting standard. If an inspection report contains a keyword related to a particular

accounting standard, this indicates the auditor had a deficiency related to auditing this standard. We also

determine how frequently the accounting standard keywords are used in each client’s financial statements

to understand the relative exposure each client has to the accounting standard. We then examine how

clients’ exposure to accounting standards related to areas where their auditor has PCAOB identified

deficiencies affects changes in audit fees and auditor changes.

We find that when a client-auditor pair has a high level of exposure to deficient auditing, based

on the client’s relative use of the accounting standards and the auditor’s deficiencies in these areas

relative to other Big 4 auditors, there is a higher likelihood of an auditor change occurring. There is also

lower deficient auditing exposure in the resulting new auditor-client pair. Deficient auditing exposure,

however, does not appear to be related to changes in audit fees for auditor-client relationships that are

maintained. Together, these results suggest that information in PCAOB inspection reports affect auditor-

client relationships by helping to align clients’ accountings with Big 4 auditors’ expertise.

This paper makes a number of contributions to existing literature. First, the paper further

explores the effects of PCAOB inspections on auditor-client relationships. The literature to date has

found only limited evidence that PCAOB inspections affect these relationships. Lennox and Pittman

(2010) conclude clients do not find the inspection reports informative because auditors’ quality control

system deficiencies are not publically disclosed. Other studies provide evidence that triennially inspected

auditors (auditors with fewer than 100 clients) are more likely to be dismissed when they have GAAP-

deficient PCAOB reports (Abbott, Gunny, and Zhang 2013) and that smaller, lower quality auditors exit

4

the audit market (Defond and Lennox 2011). Our study, however, focuses on the effects that PCAOB

inspection reports have on Big 4 auditors and their clients. Included in our study is an examination of the

reports’ effects on audit fees, an aspect of the auditor-client relationship that has received little attention

in prior studies on the effects of PCAOB inspection reports.

Second, our paper introduces a new approach to examining deficiencies identified in PCAOB

inspection reports. Specifically, we link PCAOB inspection deficiencies to firms’ financial statements

using an accounting standard keyword list developed by Folsom, Hribar, Mergenthaler, and Peterson

(2012). Prior work has relied on the number of issues identified in PCAOB inspection reports (e.g.,

Lennox and Pittman 2010; Abbott, Gunny, and Zhang 2013), which may be less relevant to clients’

auditor-engagement decisions than the types of deficiencies found. Our method is particularly

advantageous because the PCAOB releases little information about the clients where inspections issues

are identified.3 Because the client or the industry in which the client operates cannot usually be

identified, linking deficiencies related to accounting standards related with clients’ exposure to those

standards may be the best option for determining how specific types of PCAOB audit deficiencies affect

clients.

Our paper also has implications for auditing regulation and practice. Our results on PCAOB

inspection reports and audit fees provide information on whether the inspection reports improve audit

quality. We would expect fees to increase for clients with exposure to accounting standards related to

PCAOB audit deficiencies if auditors begin to do more testing to remediate these deficiencies. We would

expect lower fees if auditors reduce fees to pacify their clients after negative PCAOB reports. Lower fees

could hinder resolving the audit deficiencies and may lead to deficiencies in other areas if the auditor is

spread thin as a result of the audit fee cuts. Our results do not show evidence of either scenario, which

suggests that Big 4 audit firms with more deficiency exposure have little ability to negotiate higher fees,

but also do not appear to lower fees to retain clients. Further, our results provide evidence on PCAOB

3 Language in the SOX legislation actually prohibits the PCAOB from releasing the names of the clients where the inspection

issues are identified.

5

inspections improving audit quality by aligning clients with auditors that have more expertise in the

accounting standards used by the clients. This evidence has important implications for proposed rules

requiring auditor rotation. Auditor rotation could possibly have a negative effect on audit quality because

it could undo gains related to auditor-client realignment resulting from PCAOB inspections.

2. Background and Hypothesis Development

2.1 Background on PCAOB inspection process

The Public Company Accounting Oversight Board (PCAOB) was established by the Sarbanes-

Oxley Act of 2002 (SOX) as part of sweeping reforms aimed at restoring investor confidence in capital

markets after a number of high-profile accounting frauds. The PCAOB is charged with overseeing public

company auditors by establishing auditing standards, performing inspections, and assessing penalties for

failure to adhere to its standards. Prior to the establishment of the PCAOB, the audit profession was

largely self-regulated. The American Institute of Certified Public Accountants (AICPA) created auditing

standards and established a peer review process in which member firms selected another audit firm to

perform an annual review.4

The Sarbanes-Oxley Act requires audit firms that audit over 100 public clients to be inspected

annually and auditors with fewer public clients to be inspected every three years (PCAOB 2008). The

PCAOB inspections program uses a risk-based approach to selecting audit engagements, and audit areas

within those engagements, and thus the inspection findings are not necessarily representative of all audit

engagements (Gradison and Boster 2010). Inspectors may identify audit deficiencies as well as GAAP

departures, and these findings are made public in a report released typically several months following the

inspection (although the client is not identified in the public report). Quality control deficiencies

identified during the inspection process are only made public if they are not remediated to the Board’s

satisfaction within 12 months.

4 While this peer review process was self-regulated, there is some evidence that it provided audit clients with useful

information. Hilary and Lennox (2005) find that audit firms gained and lost clients depending on whether there were

positive or negative peer review opinions.

6

2.2 Impact of the PCAOB Inspection Process

To date, there is limited evidence on the impact of the PCAOB inspection process on audit

quality. In fact, the PCAOB and others have acknowledged it is difficult to measure “audit quality,”

which in turn makes it difficult to assess improvements in audit quality (e.g. Department of the Treasury,

2008 p. VIII:14-15). Carcello, Hollingsworth, and Mastrolia (2011) provide evidence related to audit

quality, as proxied by changes in abnormal accruals, for clients of the Big 4 audit firms following the first

two years of inspections. Carcello, Hollingsworth, and Mastrolia (2011) document a significant reduction

in abnormal accruals and conclude their results are consistent with an improvement in audit quality for the

Big 4 firms following PCAOB inspections. Lamoreaux (2013) compares audit quality, as proxied by the

propensity to issue going concern opinions as well as disclose material weaknesses in internal controls, in

foreign jurisdictions subject to PCAOB inspections versus those that have refused the PCAOB access to

inspect, and finds evidence of higher audit quality in those jurisdictions with “inspection exposure.” He

finds no evidence of such a difference prior to the PCAOB’s existence, and concludes exposure to the

PCAOB inspection process has resulted in an improvement in audit quality. Similarly, Gramling,

Krishnan and Zhang (2011) find an increased likelihood of issuing going concern opinions following

PCAOB inspections of triennially-inspected audit firms.

Evidence related to audit quality may also be inferred by the findings form DeFond and Lennox

(2011) who document that more than 600 small audit firms ceased auditing public clients after the

passage of SOX. DeFond and Lennox (2011) note that these were lower quality auditors since they either

avoided AICPA peer review or PCAOB inspection or were more likely to have deficiencies if they did

have a review or inspection (DeFond and Lennox 2011). The publicly-traded client firms would then be

forced to switch to an audit firm still registered with the PCAOB, presumably a higher quality audit firm.

The authors do in fact find support that audit quality, measured by the likelihood of receiving a going

concern opinion, increased for these client firms following the switch.

In this study, we seek to provide evidence on how the inspection process itself, and in particular

the PCAOB’s focus on high risk audit areas, may improve (or possibly reduce) audit quality. While prior

7

studies provide evidence of audit quality on average, as proxied by either the likelihood of issuing a going

concern opinion or abnormal accruals, we examine whether audit firms and clients respond to the

particular audit and/or GAAP deficiencies identified in the inspection process. As discussed earlier, the

PCAOB’s focus on particular audit areas will likely impact both audit firm behavior and client firms’

assessments of their current auditor. The inspection process sheds light on audit deficiencies and audit

firms may increase audit effort in these areas, or may even choose to avoid client firms with significant

exposure to particular high risk areas. In addition, the inspection process may provide evidence of an

audit firm’s quality which could result in client firms renegotiating audit fees or seeking a new auditor.

We discuss each of these potential impacts in greater detail below.

2.2.1 Audit Firm Responses to PCAOB Inspections

There tend to be common themes across PCAOB inspection reports in terms of the audit areas

reviewed and deficiencies identified. For example, in their summary of 2004-2007 inspection findings for

the annually inspected firms, the PCAOB notes that it continues to observe deficiencies in the significant

areas of revenue, fair value, management’s estimates, determination of materiality, and audit scope

(PCAOB 2008). The PCAOB also notes that it will continue to focus on these significant areas in future

inspections. In a study summarizing the PCAOB inspection findings for the annually inspected firms

over the period 2004-2009, Church and Shefchik (2012) also note the most common deficiencies involve

testing revenues, fair value measurements, other accounting estimates, and internal controls. Thus, the

engagements selected for review by the PCAOB are likely to be those with significant exposure to the

areas of concern. For example, the PCAOB paid particular attention to audits of companies in the

financial services sector following the financial crisis (PCAOB 2010).

There are various ways in which audit firms may respond following the identification and

disclosure of audit deficiencies. Audit firms may merely become more conservative in their reporting

decisions. For example, the findings of Lamoreaux (2013) may be evidence of changes in reporting

decisions in that audit firms increase the likelihood of issuing a going concern opinion and disclosing a

8

material weakness in internal controls. This does not necessarily imply they have changed their audit

approach to increase the amount of audit evidence gathered in these areas.

Alternatively, auditors may choose to increase audit effort in particular high risk areas they

believe the PCAOB will focus on during the inspection, or that were focus areas in previous inspections.

Carcello, Hollingsworth, and Mastrolia (2011) note that firms can adjust their audit approach, address

issues through their continuing education programs, and strengthen their internal work paper review

procedures in response to deficiencies identified in an inspection. For example, audit firms often note in

their written response to an inspection report that they are increasing training and documentation in

particular areas as a result of deficiencies identified by the PCAOB. We expect this would translate into

additional audit effort and fees for client firms with significant exposure to the audit area in which a

deficiency was identified (e.g. fair value estimates).

In some cases, an audit firm may choose to resign from clients with significant exposure in areas

where the audit firm performs poorly. When the PCAOB identifies a deficiency at an audit firm, the firm

may perceive clients with heavy exposure to the related accounting standard as higher risk because the

firm will have to make changes to its audit procedures for those clients. Further, auditors may perceive

higher reputation risk from auditing these clients because continuing to audit them may result in future

deficiencies identified by the PCAOB.

2.2.2 Client Firm Responses to PCAOB Inspection Reports

Since the passage and implementation of SOX, the audit committee has been responsible for the

hiring and firing the external auditor and for evaluating them on an annual basis. The PCAOB believes

their reports should be informative to audit committees making auditor acceptance and retention

decisions. PCAOB Chairman James Doty, for example, states, “Whether an audit committee's own

company audit is being reviewed as part of an inspection, or whether it's another company within the

same industry, PCAOB inspection reports provide insight into areas of risk and audit quality that are of

concern to all audit committees” (Whitehouse 2012).

9

In a discussion paper aimed at audit committee members, the PCAOB provides recommended

questions the audit committee should present to their audit firm in order to understand the implications of

inspection findings (PCAOB 2012). One of the issues they recommend audit committees discuss is

whether or not deficiencies identified on other audit engagements are related to audit areas of their

company. In addition, the PCAOB recommends the audit committee ask what the audit firm is doing to

remediate any identified deficiencies. This document suggests the PCAOB believes there is useful

information for audit committees in the inspection reports that should be considered in their annual

evaluation of the incumbent audit firm, even if their company’s audit was not selected for inspection.

Audit committees and management may react to this information in various ways. An optimistic

view is that they encourage their audit firm to invest in training and developing the appropriate audit

methodology to address the deficiencies, and state they support their auditor in increasing audit effort in

high risk areas. On the other hand, a pessimistic view is that the audit committee finds the inspection

reports uninformative, unrelated to their particular engagement, or dismisses the deficiencies as being

trivial and insignificant, and argues against any additional work that would increase fees.

Several studies discuss and examine the information content of PCAOB inspection reports, but

evidence on the usefulness of the reports is limited. One reason the reports may be viewed as

uninformative by audit committees and others is that the selection of audit engagements is risk-based and

therefore the findings are not generalizable or necessarily representative of audit quality for any particular

audit firm. Others suggest that the inspections focus on trivial and/or inconsequential audit deficiencies

(Glover, Prawitt, and Taylor 2009; DeFond 2010) and thus may not be informative to audit committees.

Despite these concerns, some evidence indicates that the reports are associated with audit quality or

perceptions of audit quality.

Gunny and Zhang (2012) examine whether PCAOB inspection reports provide information about

audit quality by associating the severity of deficiencies with proxies for audit quality. They find the

inspection reports for triennially-inspected firms provide information about audit quality, i.e. clients of

auditors with “seriously deficient” reports (reports that claim deficiencies lead to GAAP departures) have

10

significantly higher abnormal accruals and are more likely to restate their financial statements. However,

this report classification does not distinguish among the annually inspected audit firms, as there is less

variability in these reports in terms of the number and severity of deficiencies. In addition, Abbott,

Gunny, and Zhang (2013) find that clients of triennially inspected auditors with GAAP deficiencies

identified in their PCAOB inspection reports are more likely to dismiss their auditors, and this association

is stronger for companies with greater agency conflicts and more effective audit committees. The Abbott,

Gunny, and Zhang (2013) results suggest clients of triennially inspected audit firms use the inspection

reports as a signal of audit quality.

In contrast, Lennox and Pittman (2010) conclude that less information about audit quality is

conveyed with the inception of the PCAOB inspection process, as compared to what was conveyed under

the prior peer review regime, and that audit clients do not find the PCOAB inspection reports useful in

signaling audit quality. Their conclusion is based on a lack of association between the number (or

unexpected number) of deficiencies disclosed in PCAOB inspection reports and changes in audit firm

market shares. Lennox and Pittman (2010) also note there is criticism over the initial confidentiality of

quality control concerns (Part II of the PCAOB reports), and suggest this lack of transparency contributes

to the lack of informativeness of the reports. However, as noted by DeFond (2010), even if the reports are

“uninformative” to client firms, the inspection process may still result in improvements in audit quality.

In a more recent study examining the public disclosure of Part II findings for triennially inspected firms

that do not remediate quality control deficiencies to the satisfaction of the PCAOB within 12 months,

Ragothaman (2012) finds the PCAOB reports disclose a greater number of quality control deficiencies

relative to peer review reports, suggesting the PCAOB is more harsh than the peer reviewers and may

provide an ex ante incentive to improve audit quality.

2.2.3 Types of Deficiencies and the Audit Market

It is important to understand how audit firms and their clients react to the deficiencies as this has

implications for audit market concentration and audit quality. In the final report from The Department of

the Treasury Advisory Committee on the Auditing Profession, the committee notes “Auditing firms,

11

public companies, market participants, academics, investors and others reasoned that large public

companies with operations in multiple countries need auditing firms with global resources and technical

and industry expertise to deal with an increasingly complex business and financial reporting

environment” (Department of the Treasury 2008, p. VIII:2). The committee concludes that these needs

restrict the choice of auditor for the large multinational companies to just the global auditing firms. As

accounting standards increase in complexity and the PCAOB focuses on the auditing of these most

complex standards, more clients may be driven to seek the expertise of a global auditing firm, and even

limit their choices among the Big 4 accounting firms.

At the same time, regulators express concern that too much concentration in the audit market

decreases quality-based competition and may result in reduced audit quality (Department of the Treasury

2008, p. VIII:2-3). To the extent we observe auditor-switching activity, rather than increasing audit fees

related to improved audit procedures in response to deficiencies identified in PCAOB inspection reports,

reduced audit market competition could be an unintended consequence of the inspection process.

2.3 Hypotheses

There is little prior research on how auditors and clients react to PCAOB inspection reports and

how these reports affect auditor-client relationships. While there is some evidence that clients are more

likely to dismiss smaller, triennially-inspected audit firms with inspection deficiencies, there is no such

evidence for clients of Big 4 auditors. An important question is whether audit deficiencies affect auditor-

client relationships in terms of audit fees charged and the engagement decisions of both clients and

auditors, especially for Big 4 auditors who audit the majority of large publically traded companies.

As noted above, audit firms may respond to PCAOB inspection findings by increasing training

and audit effort in the specifically-identified areas of concern, and may face more pressure to do so if the

deficiencies are specific to the auditor as opposed to deficiencies identified for all auditors. We expect

this would result in higher audit effort and higher audit fees for clients with significant exposure to the

accounting standards related to the identified audit deficiency, leading to our first hypothesis. It is also

possible, however, that audit firms may decrease their fees in order to pacify clients considering

12

dismissing auditors that have deficiencies identified in areas important to those clients (Abbot, Gunny,

and Zhang 2013). Such a fee reduction, is also more likely to occur if the deficiencies are specific to the

auditor.

Hypothesis 1: Audit fees are increased (reduced) when a Big 4 auditor’s inspection report

indicates the exposure to deficient auditing, given the client’s use of accounting standards, is

higher than it would be for other Big 4 auditors.

As an alternative to increased audit effort, audit firms may selectively resign from audit

engagements or price themselves out of the market to avoid significant exposure to particular high risk

audit engagements. Client firms may also choose to dismiss their audit firm if they believe they do not

have sufficient expertise in auditing certain high risk areas. In both cases, we expect an increase in

switching activity away from an audit firm without expertise to an audit firm perceived to have greater

expertise. These predictions lead to the following two hypotheses.

Hypothesis 2a: An auditor-client engagement is less likely to continue when a Big 4 auditor’s

inspection report indicates the exposure to deficient auditing, given the client’s use of accounting

standards, is higher than it would be for other Big 4 auditors.

Hypothesis 2b: On average, clients switching Big 4 auditors is will engage an auditor that results

in lower exposure to deficient auditing than existed with the prior auditor.

3. Research Design

3.1 Data

We begin our analysis with PCAOB inspection reports for the Big 4 auditing firms between 2005

and 2011.5 The PCAOB conducts annual inspections of these of these auditing firms, resulting in 28

inspection reports. The reports issued in a given year are usually based on inspections that occur in the

prior year and are assumed to be related to audits of the clients’ fiscal years ending the year prior to the

inspection work.

5 All PCOAB inspection reports are available at PCAOB.org. Full PCAOB inspections on Big 4 auditors are first

issued in 2005. In 2004, the PCAOB issued limited inspection reports on these firms, reviewing 16 audits at each

firm, whereas recent inspections typically review more than 50 audits. To the extent audit fee changes or auditor

changes occurred due to these limited reports, we expect not including the limited reports to bias against identifying

a relation between PCAOB reports and audit-client relationships.

13

Part I.A of the inspection reports, “Review of Audit Engagements,” includes descriptions of the

audit deficiencies identified by the PCAOB during their inspections. For this section in each inspection

report, we use a keyword list for U.S. GAAP standards developed by Folsom, Hribar, Mergenthaler, and

Peterson (2012) to identify accounting standards related to audit deficiencies detailed by the PCAOB.6

An auditor is considered to be “deficient” in auditing accounting related to a standard if a keyword related

to the accounting standard is identified in the inspection report. Whether a deficiency related to a certain

accounting standard is identified is a function not only of the auditing proficiency, but also a function of

the extent to which its clients use the accounting standard and the PCAOB’s focus on the area. However,

we believe that given Big 4 firms have broad client bases, and given that the PCAOB uses a risk-based

approach to selecting the audits it reviews and its focus on certain accounting issues, the likelihood of the

PCAOB reviewing auditing related to the accounting standards is relatively consistent across firms.

Table 1 displays the number of Big 4-firm inspection reports with deficiencies related to each

accounting standard that were identified using the keyword list.7 We find that standards related to

intangible assets (SFAS 142), business combinations (SFAS 141), and income taxes (SFAS 109) are most

commonly identified in the reports. These findings are aligned with the PCAOB’s 2008 report, which

stated that inspection issues were frequently identified relating to accounting estimates, fair value

estimates, and income taxes. Inspections also frequently identified issues related to revenue, which

corresponds to standards SOP 97-2, SAB 101, and SFAS 48 in the key word list (PCAOB 2008). In

addition, the PCAOB identifies inventory, fair value of financial instruments, and valuation of pension

plan assets as areas where audit firms frequently failed to test controls (PCAOB 2010). Keywords for

6 Appendix A provides a partial list of the keywords used and the related standards. See Folsom, Hribar,

Mergenthaler, and Peterson (2012) for complete details on the development of the keyword list. Folsom, Hribar,

Mergenthaler, and Peterson (2012) validate the list’s ability to identify accounting standard usage in 10-K filings by

examining the word counts across industries, correlating the word counts with dollar amounts, and having the list

reviewed by the national office of a Big 4 accounting firm. Appendix B provides the details of selected deficiencies

identified by the PCAOB. 7 Results for certain standards that contain overlapping terms are removed to avoid double counting. For example,

we include the results for search terms related to SFAS 123R, but not related to SFAS 123 and ABP 25.

14

SFAS 107, SFAS 115, SFAS 133, and SFAS 87 are related to the fair value of financial instruments and

pensions, however, our keyword list does not identify any deficiencies related to inventory.

We use the same keyword list to identify the relative importance of each accounting standard to

each audit client. Specifically, we match the keyword list to Big 4 auditor clients’ machine-readable 10-K

filings on EDGAR and count the number of times keywords related to each standard are identified in each

10-K filing. These counts are then standardized, by subtracting the mean and dividing by the standard

deviation of the counts for each standard-year, to create a measure of the relative impact (RELIMPACT)

that each standard has for a client (Folsom, Hribar, Mergenthaler and Peterson 2012). The minimum

standardized count value for each accounting standard is added back to the variable to keep all values

positive.

We match relative impact numbers from the 10-K filings to the auditor deficiencies identified in

the PCAOB inspection reports using the fiscal year of the reviewed audits and the companies’ auditors

listed on Compustat. The combination of this data allows us to calculate the amount of exposure there is

to auditing considered deficient by the PCAOB for each client-auditor pair. Specifically, we measure

deficient auditing exposure (DEFEXP) as

∑

where

RELIMPACT = client’s standardized keyword count for an accounting standard;

DEFICIENCY = indicator variable equal to one if the PCAOB identifies a client’s auditor as

having a deficiency related to the accounting standard, and zero otherwise.

To measure deficient auditing exposure with the current auditor relative to the exposure to

deficiencies they would face with other auditors, we calculate an adjusted auditor deficiency exposure

(ADJEXP) as

where

15

ALTEXP = the mean of deficient auditing exposures calculated as if the client was instead

audited by the other three Big 4 auditors.

This adjustment to the deficient auditing exposure ensures the variable captures the auditing deficiency

exposure that is specific to the auditor-client pair as opposed to exposure to auditing related to accounting

standards that is particularly high risk or a particular focus of the PCAOB. The adjusted deficiency

exposure (ADJEXP) can be thought of as the misalignment between a client’s accounting standard usage

and the auditor’s ability to audit those accounting standards.

Figure 1 provides time trends for the adjusted deficiency exposure (ADJEXP) and its components

over the fiscal years of the audits inspected by the PCAOB. In general, both adjusted deficiency

exposure and the raw values of deficiency exposure (DEFEXP) are decreasing over time. These trends

indicate that number of deficiencies is decreasing and that alignment between clients’ accounting and

auditor expertise is improving. The increase in adjusted deficiency exposure and its components in 2007

is primarily driven by one auditor’s inspection, which contained new deficiencies related to taxes and

intangible assets.

Other data for the models is obtained from Audit Analytics and Compustat. Data on audit fees

and auditor changes are from Audit Analytics. Client data for control variables in the models is obtained

from Compustat. We require sample observations to have all necessary data items for the models and

exclude financial firms (SIC 60-69) because these clients are likely to have different audit fee structures

from other clients (Fields, Fraser, and Wilkins 2004).

3.2 Models

3.2.1 Audit Fee Changes

We investigate Hypothesis 1 on whether changes in deficient auditing exposure affects audit fees

using a model of audit fee changes. Specifically, we use an ordinary least squares (OLS) regression of

audit fee changes on changes in deficient auditing exposure and control variables. The variables in the

model are based primarily on work by Ettredge, Li, and Scholz (2007) and Hogan and Wilkins (2008).

The form of the model is

16

where

ΔFEEit = change in the natural logarithm of audit fees for client i from fiscal year t-1 to fiscal

year t;

ΔADJEXPit = change in adjusted auditing exposure between the current PCAOB report issued

during fiscal year t and the last PCAOB report prior to start of fiscal year t for client i’s auditor;

ΔSIZEit = change in the natural logarithm of total assets for client i from fiscal year t-1 to fiscal

year t;

ΔROAit = change in return on assets for client i from fiscal year t-1 to fiscal year t;

ΔABSAACCit = change in the absolute value of abnormal accruals for client i from fiscal year t-1

to fiscal year t;

ΔNSOPit = change in an indicator variable for a non-standard audit opinion for client i from fiscal

year t-1 to fiscal year t (1 indicates change from standard to non-standard, -1 indicates change

from non-standard to standard, 0 indicates no change);

ΔMWit = change in an indicator variable for the existence of an internal control material

weakness for client i from fiscal year t-1 to fiscal year t (1 indicates change from no weakness to

weakness, -1 indicates change from weakness to no weakness, 0 indicates no change);

ΔLEVit = change in leverage for client i from fiscal year t-1 to fiscal year t;

ΔINVit = change in inventory scaled by total assets for client i from fiscal year t-1 to fiscal year t;

ΔRECit = change in receivables scaled by total assets for client i from fiscal year t-1 to fiscal year

t;

ΔM&Ait = change in an indicator variable for mergers and acquisition activity for client i from

fiscal year t-1 to fiscal year t (1 indicates change from no M&A activity to M&A activity, -1

indicates change from M&A activity to no M&A activity, 0 indicates no change);

ΔSEGSit = change in the number of reported business segments for client i from fiscal year t-1 to

fiscal year t;

ΔSPITEMit = change in an indicator variable for the existence of special items for client i from

fiscal year t-1 to fiscal year t (1 indicates change from no special items to special items, -1

indicates change from special items to no special items, 0 indicates no change);

ΔRESTRUCTit = change in an indicator variable for restructuring charges for client i from fiscal

year t-1 to fiscal year t (1 indicates change from no restructuring charges to restructuring charges,

-1 indicates change from restructuring charges to no restructuring charges, 0 indicates no change);

YEAR = indicator variables equal to one if the observation for the corresponding fiscal year, and

zero otherwise.

Complete variable definitions are available in Appendix C.

The variable of interest, ΔADJEXP, measures the change in deficient auditing exposure for the

client-auditor pair, relative to the average exposure that would occur if the client used other Big 4

auditors. A positive coefficient on ΔADJEXP would indicate that audit firms increase their fees, either

due to increased risk or increased audit effort, when the auditor-client relationship experiences an increase

in exposure to deficient auditing. A negative coefficient would be consistent with auditors attempting to

17

pacify clients that might otherwise dismiss the auditor and replace them with an auditor with fewer audit

deficiencies in areas that are important to the client.

We control for a number of other variables we expect to be related to changes in audit fees from

one year to the next. We expect that increases in firm size (ΔSIZE), inventory (ΔINV), receivables

(ΔREC), and business segments (ΔSEGS) will be positively related to changes in audit fees because these

variables are related to audit effort. Similarly, additional audit effort is likely to increase audit fees when

clients engage in more special transactions such as restructuring (ΔRESTRUCT), mergers and acquisitions

(ΔM&A), or special items (ΔSPITEM) or when clients have new auditing issues that lead to non-standard

audit opinions (ΔNSOP) or the disclosure of material weaknesses in internal control (ΔMW) (Ettredge, Li,

and Scholz 2007; Hogan and Wilkins 2008). We also control for variables related to audit risk that will

increase audit effort or the premium the auditor charges for bearing the risk. Changes in audit fees are

expected to be negatively related to changes in return on assets (ΔROA) and positively related to changes

in the absolute value of abnormal accruals (ΔABSAACC) and changes in leverage (ΔLEV). Finally, we

control for the client fiscal YEAR to the capture time trend in audit fee changes.

3.2.2 Auditor Changes

We use a logistic regression model to investigate Hypothesis 2 on whether the adjusted auditing

deficiency exposure affects the likelihood of auditor changes occurring. The unit of analysis is a client-

inspection observation. For each annual PCAOB inspection report (r), there is one client-inspection

observation for each publically traded client of the audit firm that has the necessary data. The dependent

variable (AUDCH) is equal to one if an auditor change occurs between the inspection report release date

and the lesser of one year or the auditor’s next inspection report release, and zero otherwise. The control

variables in the model are based on Ettredge, Li, and Scholz (2007). The form of the model is

where

18

AUDCHir = indicator equal to one if client i changes auditors between the time its auditor’s

inspection report is released and the lesser of one year or the auditor’s next inspection report

release, and zero otherwise;

ADJEXPir = adjusted auditing exposure calculated for client i using inspection report r;

SIZEir = natural logarithm of total assets for client i for the fiscal year ending prior to the release

of inspection report r;

ABFEEir = abnormal audit fees for client i for the fiscal year ending prior to the release of

inspection report r;

NSOPir = indicator variable equal to one if client i has a non-standard audit opinion for the fiscal

year ending prior to the release of inspection report r, and zero otherwise;

MWir = indicator variable equal to one if client i has an internal control material weakness for the

fiscal year ending prior to the release of inspection report r, and zero otherwise;

LEVir = leverage for client i for the fiscal year ending prior to the release of inspection report r;

ABSAACCir = absolute value of abnormal accruals for client i for the fiscal year ending prior to

the release of inspection report r;

AUDINDSHRir = percent of the square root of client assets audited by client i’s auditor in client

i’s two-digit standard industry code for the fiscal year ending prior to the release of inspection

report r;

INDUSTRY = indicator variables equal to one if firm i is a member of the two-digit Standard

Industry Classification (SIC) code, and zero otherwise;

YEAR = indicator variables equal to one if the observation for the corresponding fiscal year, and

zero otherwise.

Complete variable definitions are available in Appendix C.

In this model, the variable of interest, ADJEXP, is the auditing deficiency exposure for the client-

auditor pair less the average exposure that would have occurred if the client used other Big 4 auditors.

We expect a positive relation between ADJEXP and auditor changes, which would indicate that auditor

changes are more likely to occur when auditors have deficiencies related to areas of accounting important

to the client. The client may be more likely to dismiss the auditor because other auditors would have

fewer deficiencies or the auditor may be more likely to resign because the client is seen as being higher

risk to the auditor.

Other variables likely to affect auditor-client engagement decisions are included in the model as

controls. We expect SIZE to be negatively related to auditor changes (AUDCH), because large clients are

likely to be important to the audit firms. We also expect the auditor’s market share of the industry

(AUDINDSHR) to be negatively related to auditor changes because both the auditor and the client benefit

from the auditor’s industry expertise. Positive coefficients are expected on leverage (LEV) and abnormal

accruals (ABSAACC) because higher values of these variables are associated with more risk, which will

19

increase the likelihood of auditor resignations. Abnormal fees (ABFEE) are expected to have a positive

relation with auditor changes because higher abnormal fees may be a sign that the auditor considers the

client to be high risk and because clients are more likely to leave auditors charging high fees (Hribar,

Kravet, and Wilson 2010; Ettredge, Li, and Sholz 2007). Further, we expect non-standard audit opinions

(NSOP) and material weakness disclosures (MW) to be positively related to auditor changes because these

variables indicate higher risk and auditor-client tension. Finally, we include indicator variables to control

for INDUSTRY and fiscal YEAR effects on auditor changes.

4. Results

4.1 Audit Fee Changes

Table 2 provides descriptive statistics for the audit fee changes test. The sample is comprised of

year-to-year audit fee changes for Big 4 auditor clients that maintain the same auditor between the years.

The sample is further limited to observations where a PCAOB inspection report is issued for the client’s

auditor between the years. These data restrictions and data requirements to the model yield a sample of

6,769 audit fee change observations. The “average” client firm is charged about $2.6 million in audit fees

and the overall trend in fees (ΔFEE) is flat over the sample period. Further, the average firm in the

sample is increasing in size (ΔSIZE), decreasing in return on assets (ΔROA), and is also decreasing in

adjusted auditing exposure (ΔADJEXP), and the likelihood of having a non-standard audit opinion

(ΔNSOP) or a material weakness in internal control (ΔMW).

In Table 2 Panel B, we separate the sample into observations with and without increases in audit

fees. Audit fees increase in 3,429 observations or approximately 51 percent of the sample. There is only

a weakly significant (Z-Stat = 1.68) difference in the median change in adjusted auditing exposure

(ΔADJEXP) between client firms with and without audit fee increases, and no significant difference in

mean values. Client firms experiencing increases in audit fees tend to be increasing in size (ΔSIZE), and

have increases in receivables (ΔREC) and inventory (ΔINV). Clients with increasing audit fees also tend

to have increasing leverage (ΔLEV), decreasing ROA (ΔROA), increasing absolute value of abnormal

20

accruals (ΔABSAACC), and new non-standard audit opinions (ΔNSOP) and material weaknesses (ΔMW).

Further, these client firms also tend to have new mergers and acquisition activity (ΔM&A) and new

special items (ΔSPITEM). Surprisingly, however, increasing audit fees appear to be related to a decline in

restructuring rather than an increase in it (ΔRESTRUCT).

Results for the audit fee changes regression are presented in Table 3. The coefficient on the

change in adjusted deficient auditing exposure (ΔADJEXP) is insignificant, indicating a lack of evidence

for changes in audit fees related to exposure to deficient auditing for clients that do not change auditors.

This result suggests that auditors with deficiencies have difficulty increasing audit fees to cover the cost

of additional audit procedures needed to remediate their deficient auditing, but also suggests auditors do

not cut fees in order to retain clients where there is high exposure to deficient auditing.

Other significant variables in the model have the predicted signs. We find audit fee increases

(ΔFEE) are positively related to increases in size (ΔSIZE), receivables (ΔREC), inventory (ΔINV),

leverage (ΔLEV), and the absolute value of abnormal accruals (ΔABSAACC), and negatively related to

return on assets (ΔROA) consistent with these variables increasing audit risk and effort. Further, increase

in audit fees are related to new mergers and acquisitions activity (ΔM&A), new special items (ΔSPITEM),

and new material weaknesses in internal control (ΔMW), which indicates these items also increase audit

risk and fees.

4.2 Auditor Changes

Descriptive statistics for the auditor changes test are provided in Table 4. Limiting the sample to

Big 4 clients and data requirements yields a sample of 11,359 client-inspection report observations. The

inspection reports lead to auditor changes within the lesser of one-year and the next inspection report in

456 observations or four percent of the sample. On average, the client firms in the sample have total

assets (SIZE) of approximately $700 million and long-term-debt-to-asset ratios (LEV) of slightly less than

0.2.

Table 4 Panel B separates the sample into observations with and without auditor changes. The

two groups are significantly different in mean and median for all variables in the auditor change model

21

except for the likelihood of non-standard audit opinions (NSOP) and the auditor’s market share of the

audit clients in the industry (AUDINDSHR). Importantly, the auditor switching observations have

significantly higher values of adjusted audit deficiency exposure (ADJEXP) than observations where the

auditor-client relationship is maintained. The clients switching auditors also are smaller companies

(SIZE) with higher leverage (LEV), higher abnormal audit fees (ABFEE) more material weaknesses in

internal control (MW), and a larger absolute value of abnormal accruals (ABSAACC).

Table 5 displays the results of a logistic regression where the dependent variable equal to one if

there is an auditor switch, and zero otherwise. The coefficient on adjusted deficient auditing exposure

(ADJEXP), the variable of interest, is positive and significant at p < 0.05. This result indicates that

PCAOB reports affect auditor-client relationships by leading to more auditor changes among auditor-

client pairs where exposure to PCAOB-identified auditing deficiencies is high. This finding is in contrast

to the conclusions of Lennox and Pitman (2010) who find that the overall number of issues identified in

PCAOB inspection reports is not related to changes in auditor market share. Together, the results suggest

that while PCAOB inspections do not change overall auditor market share, they do play a role in aligning

clients’ accounting with auditors’ abilities.

A number of the control variables also have significant coefficients. Consistent with expectations

there is a significantly positive relation between the likelihood of auditor changes and both non-standard

audit opinions (NSOP) and the disclosure of material weaknesses in internal control (MW). These

relations are consistent with auditors resigning from higher risk clients and clients dismissing auditors in

search of more favorable audit reports. We also find significantly negative relations between auditor

changes and SIZE and the auditor’s market share of the client’s industry (AUDINDSHR). These results

are consistent with auditors wanting to retain their larger clients and with both auditors and clients

wanting to maintain relationships where the auditor has more specialization in the industry.

Table 6 examines how deficient auditing exposure (DEFEXP) changes for clients that switch

from one Big 4 auditor to another. Within the Big 4, auditor changes occur in 157 cases or about 34% of

22

our auditor change observations (with the remaining changes in our sample being from a Big 4 firm to a

non-Big 4 firm). In these cases, the average deficient auditing exposure (DEFEXP) significantly changes

from 6.44 before the switch to 5.83 after the switch (t-stat = -2.63). This result indicates that the new

auditor-client pairings established from an auditor change result in lower deficient auditing exposure,

which is one manner in which deficient auditing exposure is reduced overall.

Overall, the results show that information in PCAOB inspection reports does affect auditor-client

relationships in the Big 4 audit market. Specifically, auditor changes are more likely when an auditor-

client pairing creates more exposure to deficient auditing than the average pairing of the client with other

Big 4 auditors. Further, when a new auditor-client pair is established, it has lower deficient auditing

exposure than the prior pairing.

5. Conclusion

PCAOB inspection reports are highly visible critiques of auditors’ work, but many have

questioned whether the reports contain useful information. Our study investigates whether inspections

affect auditor-client relationships by examining the relation between inspection findings and audit fee

changes and auditor changes for Big 4 auditors and their clients. We examine this relation using a new

measure of the information in the inspection reports that accounts for both clients’ use of accounting

standards and their auditors’ deficiencies related to these standards compared to other Big 4 firms’

deficiencies.

We find that our measure of relative exposure to deficient auditing is positively related to auditor

changes, but is not related to changes in audit fees. These results suggest PCAOB inspections affect

auditor changes, but auditors do not have the ability to increase fees to remediate deficient auditing. It

also does not appear, however, that auditors reduce fees to retain clients when they have more

deficiencies in areas important to the clients. Our findings are in contrast to Lennox and Pittman (2010),

who find no relation between the number of deficiencies identified and changes in auditor market share.

23

Our study has import implications for auditing regulation and practice. Our results on PCAOB

inspection reports and audit fees suggest it may be difficult for auditors remediate their deficiencies since

it appears they are not able to pass higher audit costs on to clients. Our findings with respect to auditor

changes indicate that PCAOB inspections do yield useful information and that this information is used to

improve the alignment between auditor expertise and clients’ use of accounting standards. While this

trend may be beneficial in the short term, narrowing audit expertise could have a negative effect on audit

market competition in the future. Further, our study has implications for proposed rules on auditor

rotation. Such a requirement, if implemented, could have a negative effect on gains in audit quality

resulting from auditor-client realignment.

24

References

Abbott, L., K. Gunny, and T. Zhang. 2013. When the PCAOB talks, who listens? Evidence from

stakeholder reaction to GAAP-deficient PCAOB inspection reports of small auditors. Auditing: A

Journal of Practice & Theory 32(2): 1-31.

Carcello, J.V., C. Hollingsworth, and S.A. Mastrolia. 2011. The effect of PCAOB inspections on Big 4

audit quality. Research in Accounting Regulation 23 (2): 85-96.

Church, B.K. and L.B. Shefchik. 2012. PCAOB inspections and large accounting firms. Accounting

Horizons 26 (1): 43-63.

DeFond, M.L., 2010. How should the auditors be audited? Comparing the PCAOB Inspections with the

AICPA Peer Reviews. Journal of Accounting & Economics 49 (1/2): 104-108.

DeFond, M.L. and C. Lennox. 2011. The effect of SOX on small auditor exits and audit quality. Journal

of Accounting & Economics 52 (1): 21-40.

Department of the Treasury. 2008. Advisory Committee on the Auditing Profession. (October 6).

Ettredge, M.L., C. Li, and S. Scholz. 2007. Audit Fees and Auditor Dismissals in the Sarbanes-Oxley Era.

Accounting Horizons 21 (4): 371-386.

Farrell, J., and H. Shabad. 2005. The focus of future PCAOB auditor inspections. CPA Journal 75 (6), 9.

Fields, L.P., D. Fraser, and M.S. Wilkins. 2004. An investigation of the pricing of audit services for

financial institutions. Journal of Accounting and Public Policy 23(1): 53-77.

Folsom, D., P. Hribar, R. Mergenthaler, and K. Peterson. 2012. Principles-Based Standards and Earnings

Attributes. Working Paper, University of Iowa.

Glover, S.M., D.F. Prawitt, and M. H. Taylor. 2009. Auditor standard setting and inspection for U.S.

public companies: A critical assessment and recommendations for fundamental

change. Accounting Horizons 23 (2): 221-237.

Gradison, B. and R. Boster. 2010. The PCAOB’s first seven years: A retrospection. Current Issues in

Auditing 4(1): A9-A20.

Gramling, A., J. Krishnan, and Y. Zhang. 2011. Are PCAOB identified audit deficiencies associated with

a change in reporting decisions of triennially inspected audit firms? Auditing: A Journal of

Practice and Theory 30 (3)59-79.

Gunny, K.A. and T.C. Zhang. 2013. PCAOB inspection reports and audit quality. Journal of Accounting

and Public Policy 32: 136-160.

Hillary, G. and C. Lennox. 2005. The credibility of self-regulation: Evidence from the accounting

profession’s peer review program. Journal of Accounting and Economics 40: 211-229.

25

Hogan, C.E and M.S. Wilkins. 2008. Evidence on the Audit Risk Model: Do Auditors Increase Audit

Fees in the Presence of Internal Control Deficiencies? Contemporary Accounting Research 25(1):

219-242.

Hribar, P., T. Kravet, and R. Wilson. 2010. A New Measure of Accounting Quality. Working Paper, The

University of Iowa.

Lamoreaux, P.T. 2013. Does PCAOB Inspection Exposure Affect Auditor Reporting Decisions? Working

Paper, Arizona State University.

Lennox, C., and J. Pittman. 2010. Auditing the auditors: Evidence on the recent reforms to the external

monitoring of audit firms. Journal of Accounting & Economics 49 (1/2): 84-103.

Public Company Accounting Oversight Board (PCAOB). 2008. Report on the PCAOB’s 2004, 2005,

2006, and 2007 Inspections of Domestic Annually Inspected Firms. PCAOB Release No. 2008-

08. Washington DC.: PCAOB.

Public Company Accounting Oversight Board (PCAOB). 2010. Report on Observations of PCAOB

Inspectors Related to Audit Risk Areas Affected by the Economic Crisis. PCAOB Release No.

2010-006. Washington DC.: PCAOB.

Public Company Accounting Oversight Board (PCAOB). 2012. Information for Audit Committees about

the PCAOB Inspection Process. PCAOB Release No. 2012-003. Washington DC.: PCAOB.

Ragothaman, S. 2012. Does Watching the Watchdogs: An Examination of the PCAOB Quality Control

Inspection Reports on Triennially Inspected Audit Firms and the AICPA Peer Review Reports.

Working Paper, The University of South Dakota.

Whitehouse, T. 2012. PCAOB Offers Audit Committees Hints on How to Use Inspection Findings.

Compliance Week. August 1.

Appendix A. Keyword Search Terms by Standard

This appendix contains a partial listing of keyword search terms from Folsom, Hribar, Mergenthaler and Peterson (2012) used to

create DEFEXP, ALTEXP, and ADJEXP. In our textual analysis, we also include additional variations of these keywords and the

name of each standard. A complete list is available from the authors. Any keywords with “/#/” signifies the number of adjacent

words we search for relevant terms. For example, “Warrant /5/ Debt” signifies we search for keyword “warrant” within five words

(forwards or backwards) of the keyword “debt.” The “|” reflects an ‘OR’ for the associated keywords within parentheses. For example,

“(stock|share)-based compensation” would match either “stock-based compensation” or “share-based compensation”.

Standard Keyword #1 Keyword #2 Keyword #3 Keyword #4

APB 2 investment credit* /10/

tax* tax /10/ deferral method

allowable investment

credit

APB 4 investment credit* /10/

tax*

tax* /10/ flow-through

method

APB 9 extraordinary items extraordinary gain extraordinary loss

APB 14 warrant /5/ debt convertible /5/ debt stock purchase warrant* conversion option /5/

debt

APB 16 business combination merge* /5/ pool* acqui* /5/ pool merge* /5/ purchase

APB 17 goodwill intangible asset* goodwill /5/ amortiz*

APB 18 equity method significant influence share of earnings share of loss(es)

APB 20 change in accounting

principle

change in accounting

estimate change in reporting entity

error /5/ previously

issued financial statement

APB 21 non[-]interest bearing /2/

note

non[-]interest bearing /2/

receivable impute* /2/ interest note /5/ discount

APB 23 invest* /5/ permanent /5/

foreign /5/ tax

undistributed earnings /5/

subsidiar*

accounting for income

taxes /3/ special areas

unremitted earnings /5/

subsidiar*

APB 25 stock-based

compensation option(s) /5/ grant restricted stock /5/ grant option(s) /5/ issue

APB 26 early /5/ extinguish* /5/

debt

early /5/ extinguish* /5/

liabilit*

APB 29 non[-]monetary

transaction non[-]monetary exchange non[-]reciprocal transfer

APB 30 discontinued operations extraordinary items disposal /5/ segment unusual /5/ infrequent*

ARB 43 Ch. 3a current (asset|liabilit*) /5/

classif* /5/ year

current (asset|liabilit*) /5/

classif* /5/ operating

cycle

working capital

ARB 43 Ch. 3b right (of|to) setoff right (of|to) offset

ARB 43 Ch. 4 lower of cost or market

/1/ impairment inventory /1/ impairment inventory pricing

firm purchase

commitment

ARB 43 Ch. 7a quasi-reorganization corporate readjustment

ARB 43 Ch. 7b stock dividends stock split* split-ups

27

Appendix A. Keyword Search Terms by Standard (Cont.)

Standard Keyword #1 Keyword #2 Keyword #3 Keyword #4

ARB 43 Ch. 9a depreciation

ARB 43 Ch. 9b Depreciation /5/

Appreciat* Asset*

depreciation /5/

appreciat*

depreciation /15/ quasi-

reorganization

ARB 43 Ch. 10a real estate taxes property taxes real estate /5/ tax* property /5/ tax*

ARB 43 Ch. 11a cost plus /5/ fee* /5/

contract*

ARB 43 Ch. 11b government contract /5/

renegotiat*

ARB 43 Ch. 11c fixed fee /5/ war

contract* /5/ terminat*

war /5/ contract /5/

terminat*

defense /5/ contract /5/

terminat*

war and defense contract

/5/ terminat*

ARB 43 Ch. 12 foreign earnings /5/

presentation

ARB 45 percentage of completion long term construction construction /2/ progress cost /2/ excess /2/ billings

ARB 51 consolidat* /5/ financial

statement

intercompany /5/

eliminat*

controlling financial

interest minority interest

Concepts 5 & 6 earned /5/ revenue earned /5/ sales realizable future benefit probable future benefit

SFAS 2 research and develop* research /5/ cost*

SFAS 5 conting* liab* conting* gain conting* /5/ loss conting* /5/ reasonably

possible

SFAS 7 develop* stage /10/

enterpris* develop* stage /10/ corp*

develop* stage /10/

company

planned principal

operations /5/

commenced

SFAS 13 lease bargain purchase option bargain renewal option transfer* ownership /5/

lesee

SFAS 15 trouble* debt restruc* debt /5/ restruc* debt restruct /5/ settle* debt /5/ modifi*

SFAS 16 prior period adjustment

SFAS 19 exploration mineral rights proved reserves unproved reserves

SFAS 34 interest /3/ capitaliz* self-constructed asset /5/

debt

self-constructed asset /5/

interest

SFAS 35 defined benefit /5/

pension defined benefit /5/ plan

SFAS 43 compensat* absence* vacation Accru* sick accru* illness accru*

SFAS 45 franchise fee franchise /5/ sales franchise /5/ revenue

28

Appendix A. Keyword Search Terms by Standard (Cont.)

Standard Keyword #1 Keyword #2 Keyword #3 Keyword #4

SFAS 47 purchase commitment purchase obligation long term commitment long term obligation

SFAS 48 revenue /5/ right of return sales /5/ right of return

SFAS 49 sale of inventory /5/

financing arrange*

SFAS 50 record* industry music industry music /5/ advance royalty record /5/ advance

royalty

SFAS 51 cable television cable /5/ hookup

SFAS 52 reporting currency foreign currency functional currency translation adjustment

SFAS 53 motion picture license /5/ film

SFAS 57 related part*

SFAS 60 insurance contract insurance /3/ short-

duration /2/ contract*

insurance /3/ long-

duration /2/ contract* insurance /2/ claim* cost

SFAS 61 title plant

SFAS 63 broadcasting industr* network affiliation

agreement*

exhibition rights /5/

license agreement*

SFAS 65 mortgage loans mortgage-backed

securities loan fees commitment fees

SFAS 66 sale* /5/ real estate retail land sale*

SFAS 67 capitaliz* /5/ real estate

/5/ acquisition

capitaliz* /5/ real estate

/5/ develop*

capitaliz* /5/ real estate

/5/ construction

capitaliz* /5/ real estate

/5/ sale

SFAS 68 fund* /5/ research and

development

research and

development

arrangement*

research and

development /5/

obligation*

SFAS 71 cost-based rates /3/

regulat*

accounting for the effects

of certain types of

regulation

regulat* asset* regulat* liabilit*

SFAS 77 receivable /5/ recourse transfer /5/ receivable transferor /5/ receivable

SFAS 80 futures contract futures /5/ hedge

SFAS 86 technological feasibility

/5/ software

internal* /5/ develop* /5/

software cost /5/ software /5/ sold capitalized software cost

SFAS 87 Pension projected benefit

obligation

accumulated benefit

obligation funding of plan assets

SFAS 88 settlement /5/ defined

benefit (pension|plan)

curtailment /5/ defined

benefit (pension|plan)

termination benefit* /10/

employ*

29

Appendix A. Keyword Search Terms by Standard (Cont.)

Standard Keyword #1 Keyword #2 Keyword #3 Keyword #4

SFAS 97 universal life type long-duration contract

/10/ mortality

retrospective deposit

method limited payment contract

SFAS 101

discontinuation of

application of FASB

statement no. 71

cease* to meet the

criteria /5/ SFAS [No.]

71

regulated enterprise

SFAS 105 disclos* /5/ financial

instrument*

disclos* /5/ off balance

sheet risk disclos* /5/ credit risk

SFAS 106 post[-]retirement benefits

other than pensions

post[-]retirement health

care benefit

post[-]retirement benefit

plan

post[-]retirement /5/

health care

SFAS 107 disclos* /5/ financial*

instrument* /5/ fair value

SFAS 109 income tax* tax liability tax asset tax /5/ temporary

difference

SFAS 113 reinsurance retrocession

SFAS 115 available-for-sale /5/

securit* trading /5/ securit

held-to-maturity /5/

securit*

other than temporary

impairment /10/

investment

SFAS 116

accounting for

contributions received

and contributions made

nonreciprocal transfer donor imposed restriction donor imposed condition

SFAS 119 disclos* /5/ derivative disclos* /5/ futures

contract

disclos* /5/ forward

contract

disclos* /5/ swap

contract

SFAS 121 impair* /5/ long-lived dispos* /5/ long-lived

SFAS 123 (stock|share)-based

compensation

option(s) /5/

(grant*|issue*|award*)

restricted stock /5/

(grant*|issue*|award*) grant date

SFAS 123r (stock|share)-based

compensation

option(s) /5/

(grant*|issue*|award*)

restricted stock /5/

(grant*|issue*|award*) grant date

SFAS 125 transfer /5/ financ* asset* servic* /5/ financ* asset* extinguish* /2/ liabilit* transfer /5/ receivable

SFAS 130 comprehensive income

SFAS 133 derivativ* hedg* underlying /5/ notional

amount* put /2/ option

SFAS 140 transfer* /5/ financ*

asset* servic* /5/ financ* asset* extinguish* /2/ liabilit* extinguish* /2/ debt

SFAS 141 business combination merge* /5/ purchase acqui* /5/ purchase merge* /5/ contingent

consideration

SFAS 142 Goodwill intangible asset* implied fair value /5/

carrying amount impair* /5/ goodwill

SFAS 143 asset* retirement

obligation

SFAS 144 impair* /5/ long-lived dispos* /5/ long-lived

30

Appendix A. Keyword Search Terms by Standard (Cont.)

Standard Keyword #1 Keyword #2 Keyword #3 Keyword #4

SFAS 146

restruct*

(exp*|charg*|activit*|rese

rv*)

exit or disposal activit* exit /5/ disposal activity

/5/ termination benefits

SFAS 150

instruments with

characteristics of both

liabilities and equity

freestanding financial

instrument

SFAS 154 change in accounting

principle

change in accounting

estimate change in reporting entity

error /5/ previously

issued financial statement

EITF 94-03 restruct* exp* restruct* charg* restruct* activit* restruct* reserv*

EITF 00-21 revenue /10/ multiple

deliverables

multiple deliverable

arrangement*

direct cost /10/ multiple

deliverable

unit /5/ value /5/ stand

alone basis

SOP 97-2 multiple element /5/

software

objective evidence /2/

element* /5/ fair value

vendor specific objective

evidence

software /10/ revenue

recognition

SAB 101 persuasive evidence /5/

arrangement

persuasive evidence /5/

agreement delivery /5/ occur*

fee /2/ fixed /2/

determinable

31

Appendix B. Selected PCAOB Identified Audit Deficiencies

SFAS 131 – Ernst and Young 2004

In this audit, the Firm failed to identify a departure from GAAP that it should have identified and

addressed before issuing its audit report. The issuer disclosed in the notes to its financial statements two

reportable segments despite the presence of information that indicated the issuer was organized in more

than two reportable segments as defined by SFAS No. 131, Disclosures about Segments of an Enterprise

and Related Information. The improper aggregation of reporting segments resulted in the offsetting of

operating profits at one segment with losses at another.

SFAS 57 – KPMG 2005

In addition, the issuer failed to disclose in its financial statements (1) certain lease arrangements; (2)

information concerning an impaired asset and the issuer's method for determining the fair value of the

asset; and (3) information concerning related party transactions between the issuer and entities in which

its chairman and chief executive officer had an ownership interest. Omitting these disclosures was

inconsistent with, respectively, SFAS No. 13, Accounting for Leases; SFAS No. 144, Accounting for the

Impairment or Disposal of Long-Lived Assets; and SFAS No. 57, Related Party Disclosures.

SFAS 123R – PriceWaterhouseCoopers 2006

The Firm failed to test the fair value of warrants and stock-based compensation issued in two significant

transactions during the year.

SFAS 142 – Deloitte & Touche LLP 2007

The issuer performed its annual impairment test of goodwill as of an interim date and, during 2005,

elected to carry forward the fair values of all but one of its reporting units that it had used in its 2004

impairment test, as permitted under certain conditions by Statement of Financial Accounting Standards

("SFAS") No. 142, Goodwill and Other Intangible Assets ("SFAS No. 142"). The issuer asserted that the

conditions for doing so were satisfied, including the condition that the likelihood that the fair values of its

business units had declined below their book values was remote. The Firm failed to test that assertion

sufficiently (limiting its procedures to obtaining memoranda from management and making inquiries of

management), despite the following factors: (1) a decline in the issuer's market capitalization by over 25

percent between the annual goodwill impairment tests; (2) the shut down of certain operating assets and

the resulting impairment charges at one business unit; (3) the sales of certain long-lived assets during

2005; and (4) the reduction in the issuer's debt rating during 2005. In addition, despite the fact that the

issuer was exploring options to sell or spin off certain business units, the Firm failed to evaluate

sufficiently (limiting its procedures to obtaining memoranda from management and making inquiries of

management) whether the issuer's determination of the fair values of those business units appropriately

took into account whether all or any portion of those business units was more likely than not to be sold or

disposed of.

32

Appendix C. Variable Definitions

Variable Definition

Δ Change in variable from fiscal year t-1 to fiscal year t.

ABFEE Abnormal audit fees calculated as the residual from an OLS regression of the natural logarithm

of audit fees on size, sales, inventory and receivables, debt, number of business segments,

existence of a loss, existence of a going concern issue, timing of the fiscal year end and audit

work, industry, and year determinants. Data are from Audit Analytics and COMPUSTAT.

ABSAACC Absolute value of performance-adjusted abnormal accruals. Abnormal accruals are calculated

using a cross-sectional modified Jones model, estimated by two-digit SIC code. Performance

adjustment subtracts median abnormal accruals from companies in the same decile of prior-

year return on assets, two-digit SIC code, and year. Data are from COMPUSTAT.

ADJEXP DEFEXP minus ALTEXP.

ALTEXP The average of DEFEXP calculated for each of the other Big 4 accounting firms as if they were

auditing the client in the corresponding fiscal year.

AUDCH Indicator variable equal to one if the company announces a change in auditor in the fiscal year,

and zero otherwise.. Data are from Audit Analytics.

AUDINDSHR Percent of the square root of client assets audited by the firm in the two-digit standard industry

code. Data are from COMPUSTAT.

DEFEXP The sum of the products of the relative impact of each accounting standard for a company and

an indicator for whether the PCAOB finds a related issue for the company's auditor. The

relative impact of an accounting standard for a company is calculated by standardizing the

count of keywords related to the standard in the company's 10-K filing. Whether the PCAOB

finds a related issue is determined by the presence of keywords in Section I.A of the PCAOB

inspection report for the fiscal year corresponding to the 10-K filing.

FEE Natural logarithm of audit fees. Data are from Audit Analytics.

INV Total inventory scaled by total assets (INVT/AT). Data are from COMPUSTAT.

LEV Total long-term debt (DLTT) divided by total assets (AT). Data are from COMPUSTAT.

M&A Indicator variable equal to one if the company has merger and acquisition activity, as indicated

by the sales footnote (SALE_FN), and zero otherwise. Data are from COMPUSTAT.

MW Indicator variable equal to one if the company has a material weakness in internal control

disclosed as required by SOX 404, and zero otherwise. Data are from COMPUSTAT.

NSOP Indicator variable equal to one if the company has a non-standard audit opinion, and zero

otherwise. Data are from COMPUSTAT.

REC Total receivables scaled by total assets (RECT/AT). Data are from COMPUSTAT.

RESTRUCT Indicator variable equal to one if the company is undergoing restructuring, as indicated by the

disclosure of restructuring costs (RCA, RCP, RCEPS, RCD), and zero otherwise. Data are

from COMPUSTAT.

ROA Return on assets (IB/AT) for the fiscal year. Data are from COMPUSTAT.

SEGS Natural logarithm of one plus the number of business segments disclosed (BUSSEG). Data are

from COMPUSTAT.

SIZE Natural logarithm of total assets (AT). Data are from COMPUSTAT.

SPITEM Indicator variable equal to one if the company has a special item (SPI) in the fiscal year, and

zero otherwise. Data are from COMPUSTAT.

33

Figure 1. Deficient Auditing Exposure Time Trend

This figure displays the trend of deficient auditing exposure (DEFEXP), the exposure calculated if alternative Big 4 auditors

were used (ALTEXP), and the adjusted audit exposure (ADJEXP) over the client-fiscal years for which inspections were

performed. The figure uses the means the variables in each fiscal year. The figure is based on A 11,359 client-inspections for

the Big 4 audit firms.

-1.00

0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

2003 2004 2005 2006 2007 2008

DEFEXP

ALTEXP

ADJEXP

34

Table 1. Big 4-Firm Auditing Deficiencies

This table reports on the number of Big 4 firms with deficiencies related to each accounting standard for each fiscal year on

which the inspections took place. Only accounting standards related to at least one PCAOB-identified audit deficiency are

displayed in the figure.

Inspection Fiscal Year

Standard 2003 2004 2005 2006 2007 2008 Total

SFAS 142 3 4 4 3 4 2 20

SFAS 141 3 2 2 2 2 1 12

SFAS 109 2 2 2 1 2 2 11

SFAS 13 3 2 1 1 0 0 7

SFAS 133 2 0 1 1 1 2 7

SOP9 7-2 0 1 0 2 2 0 5

SFAS 5 0 0 0 1 2 1 4

SFAS 65 0 0 0 2 1 1 4

SFAS 144 1 0 1 1 0 1 4

APB 18 0 1 0 1 0 1 3

SFAS 50 0 0 1 1 1 0 3

ARB 43 Ch. 9a 2 1 0 0 0 0 3

SFAS 2 0 1 1 0 1 0 3

SFAS 107 0 0 0 1 1 1 3

SFAS 115 0 0 0 1 1 1 3

SFAS 131 2 0 1 0 0 0 3

SFAS 87 0 0 0 0 1 1 2

SFAS 123R 0 1 1 0 0 0 2

SFAS 52 1 1 0 0 0 0 2

SFAS 146 1 0 1 0 0 0 2

SAB 101 0 0 0 1 0 0 1

APB 14 0 1 0 0 0 0 1

APB 20 1 0 0 0 0 0 1

APB 23 1 0 0 0 0 0 1

SFAS 19 1 0 0 0 0 0 1

SFAS 48 0 0 0 0 1 0 1

SFAS 57 1 0 0 0 0 0 1

SFAS 86 0 0 0 1 0 0 1

SFAS 130 1 0 0 0 0 0 1

SFAS 143 0 1 0 0 0 0 1

35

Table 2. Descriptive Statistics for Audit Fee Changes

This table reports descriptive statistics for client firm-year observations used in the audit fee changes test. In Panel B, the mean t-

statistics are from two-sample t-tests and median Z-statistics are from a Wilcoxon two-sample test. Significance at the 10%, 5%,

and 1% levels are denoted *, **, and ***. All continuous variables are Winsorized at the 1% and 99% level. Variable definitions

appear in Appendix C.

Panel A: Descriptive Statistics

N Mean StDev 1Q Median 3Q

ΔFEE 6,769 0.006 0.217 -0.097 0.003 0.094

ΔADJEXP 6,769 -0.085 3.889 -2.506 -0.094 2.529

ΔSIZE 6,769 0.032 0.228 -0.057 0.035 0.125

ΔROA 6,769 -0.005 0.405 -0.038 -0.001 0.029

ΔABSAACC 6,769 0.000 0.094 -0.032 0.000 0.031

ΔSEGS 6,769 -0.015 0.806 0.000 0.000 0.000

ΔNSOP 6,769 -0.053 0.546 0.000 0.000 0.000

ΔICD 6,769 -0.013 0.224 0.000 0.000 0.000

ΔLEV 6,769 0.002 0.102 -0.024 0.000 0.017

ΔREC 6,769 0.007 0.051 -0.011 0.002 0.021

ΔINV 6,769 0.004 0.035 -0.003 0.000 0.011

ΔM&A 6,769 -0.010 0.478 0.000 0.000 0.000

ΔSPITEM 6,769 0.021 0.498 0.000 0.000 0.000

ΔRESTRUCT 6,769 0.000 0.032 0.000 0.000 0.000

Panel B: Comparison of Positive and Non-Positive Audit Fee Changes Observations

Positive Fee Change

(N=3,429)

Non-Positive Fee Change

(N=3,340) Between Sample Tests

Mean Median Mean Median Mean Median

ΔFEE 0.152 0.093 -0.145 -0.098 77.04 *** 71.24 ***

ΔADJEXP -0.021 0.051 -0.151 -0.235 1.37 1.68 *

ΔSIZE 0.075 0.063 -0.013 0.011 16.12 *** 16.64 ***

ΔROA -0.018 -0.003 0.008 0.001 -2.68 *** -5.26 ***

ΔABSAACC 0.006 0.000 -0.005 -0.001 5.12 *** 4.22 ***

ΔSEGS -0.026 0.000 -0.004 0.000 -1.11 -0.35

ΔNSOP -0.009 0.000 -0.099 0.000 6.74 *** 6.61 ***

ΔICD -0.002 0.000 -0.024 0.000 4.08 *** 4.06 ***

ΔLEV 0.009 0.000 -0.006 0.000 5.82 *** 6.25 ***

ΔREC 0.014 0.006 -0.001 0.000 11.84 *** 13.77 *** ΔINV 0.009 0.000 -0.001 0.000 12.23 *** 12.40 *** ΔM&A 0.044 0.000 -0.065 0.000 9.46 *** 9.35 *** ΔSPITEM 0.040 0.000 0.001 0.000 3.23 *** 3.25 *** ΔRESTRUCT -0.001 0.000 0.001 0.000 -1.90 * -1.88 *

36

Table 3. Model of Audit Fee Changes

This table reports the results of a client firm-year ordinary-least-squares (OLS) regression with audit fee change (ΔFEE) as the

dependent variable. The sample is comprised of 6,769 client-year observations. Significance at the 10%, 5%, and 1% levels are

denoted *, **, and ***. All continuous variables are Winsorized at the 1% and 99% level. Variable definitions appear in

Appendix C.

Parameter Prediction Coefficient t-Statistic

Intercept (B0) ? 0.070 7.47 ***

ΔADJEXP (B1) +/- 0.001 1.08

ΔSIZE (B2) + 0.183 11.09 ***

ΔROA (B3) - -0.021 -2.48 **

ΔABSAACC (B4) + 0.117 3.57 ***

ΔSEGS (B5) + 0.005 1.49

ΔNSOP (B6) + 0.008 1.55

ΔICD (B7) + 0.071 4.04 ***

ΔLEV (B8) + 0.103 3.29 ***

ΔREC (B9) + 0.261 3.74 ***

ΔINV (B10) + 0.457 5.34 ***

ΔM&A (B11) + 0.031 5.14 ***

ΔSPITEM (B12) + 0.025 5.06 ***

ΔRESTRUCT (B13) + -0.104 -1.36

Year Fixed Effects Included

N 6,769

Clusters 2,115

R2 0.20

37

Table 4. Descriptive Statistics for Auditor Changes

This table reports descriptive statistics for client firm-inspection report observations used in the audit fee changes test. In Panel

B, the mean t-statistics are from two-sample t-tests and median Z-statistics are from a Wilcoxon two-sample test. Significance at

the 10%, 5%, and 1% levels are denoted *, **, and ***. All continuous variables are Winsorized at the 1% and 99% level.

Variable definitions appear in Appendix C.

Panel A: Descriptive Statistics

N Mean StDev 1Q Median 3Q

AUDCH 11,359 0.040 0.196 0.000 0.000 0.000

ADJEXP 11,359 -0.120 3.212 -2.120 -0.147 1.643

SIZE 11,359 6.533 1.810 5.297 6.498 7.706

ABFEE 11,359 0.081 0.503 -0.255 0.082 0.408

NSOP 11,359 0.608 0.488 0.000 1.000 1.000

MW 11,359 0.061 0.239 0.000 0.000 0.000

LEV 11,359 0.187 0.211 0.001 0.136 0.290

ABSAACC 11,359 0.059 0.078 0.012 0.033 0.073

AUDINDSHR 11,359 0.247 0.085 0.190 0.234 0.299

Panel B: Comparison of Auditor Switching and Non-Auditor Switching Observations

Auditor Switch

(N=456)

No Auditor Switch

(N=10,903) Between Sample Tests

Mean Median Mean Median Mean Median

ADJEXP 0.288 0.138 -0.137 -0.159 2.77 *** 2.86 ***

SIZE 5.554 5.495 6.573 6.542 -11.86 *** -11.33 ***

ABFEE 0.131 0.153 0.079 0.080 1.84 * 2.26 **

NSOP 0.621 1.000 0.607 1.000 0.58 0.58

MW 0.147 0.000 0.057 0.000 5.35 *** 7.82 ***

LEV 0.171 0.106 0.188 0.137 1.70 * 2.17 **

ABSAACC 0.079 0.041 0.058 0.033 4.31 *** 3.67 ***

AUDINDSHR 0.240 0.228 0.247 0.235 -1.61 -1.44

38

Table 5. Model of Auditor Changes

This table reports the results of tests of a client-firm clustered logistic regression with auditor change (AUDCH) as the dependent

variable. The sample is comprised of 11,359 client firm-inspection report observations, 456 of which switch auditors

(AUDCH=1) and 10,903 of which do not switch auditors (AUDCH = 0). Significance at the 10%, 5%, and 1% levels are denoted

*, **, and ***. All continuous variables are Winsorized at the 1% and 99% level. Variable definitions appear in Appendix C.

Parameter Prediction Coefficient Chi-Square

Intercept (B0) ? 1.715 56.08 ***

ADJEXP (B1) + 0.037 4.24 **

SIZE (B2) - -0.441 285.09 ***

ABFEE (B3) + 0.113 1.20

NSOP (B4) + 0.373 11.16 ***

MW (B5) + 0.794 25.09 ***

LEV (B6) + 0.182 0.68

ABSAACC (B7) + 0.760 1.98

AUDINDSHR (B8) - -1.084 3.89 **

Industry Fixed Effects Included

Year Fixed Effects Included

N 11,359

Clusters 3,053

Psuedo-R2 0.66

39

Table 6. Auditing Exposure Change from Auditor Changes

This table reports the results of a paired t-test of differences in deficient auditing exposure before and after auditor changes. The

sample is comprised of 157 client-firms that change from one Big 4 auditor to another. Significance at the 10%, 5%, and 1%

levels are denoted *, **, and ***. Variable definitions appear in Appendix C.

Auditor-Client Pairing DEFEXP

New Pairing 5.38

Prior Pairing 6.44

Difference -1.06

Paired t-Test -2.63 **

N 157