University of Southern California Dissertations and Theses

Do shared auditors improve audit quality? Evidence from banking relationships

(USC Thesis Other)

Do shared auditors improve audit quality? Evidence from banking relationships

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

Do shared auditors improve audit quality?
Evidence from banking relationships.

Karen Ton

University of Southern California

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(BUSINESS ADMINISTRATION)
Conferred August 2015

1
ACKNOWLEDGEMENTS

I am extremely grateful to my advisor Mark DeFond for his guidance and encouragement
on this project. I also thank the other members of my committee, David Erkens, Kevin Murphy,
Mark Soliman, and Jieying Zhang, for their guidance and constructive feedback.

This paper has also benefited from discussions and comments from Kelsey Dworkis, Shane
Heitzman, Adam Johnson, Jessica Keeley, Maria Loumioti, Jeff McMullin, Paul Michas, Suresh
Nallareddy, Peter Oh, Bryce Schonberger, Biqin Xie, seminar participants at Emory University,
University of British Columbia, University of California Los Angeles, University of Connecticut,
University of Illinois at Chicago, University of Illinois at Urbana-Champaign, University of
Massachusetts at Amherst, University of Southern California, University of Oregon, Temple
University, several partners at the various Big 4 accounting firms, and several executives and
board members at various firms.

I also thank my family for their unconditional support throughout my academic career.
2
TABLE OF CONTENTS

ACKNOWLEDGEMENTS .................................................................................................................. 1
TABLE OF CONTENTS ...................................................................................................................... 2
ABSTRACT .......................................................................................................................................... 4
INTRODUCTION ................................................................................................................................ 5
LITERATURE REVIEW AND HYPOTHESIS DEVELOPMENT .................................................. 11
Background and prior literature ...................................................................................................... 11
Auditor competencies ................................................................................................................. 11
Shared Auditor ............................................................................................................................ 12
Hypothesis Development ................................................................................................................ 13
RESEARCH DESIGN AND SAMPLE.............................................................................................. 17
Sample............................................................................................................................................. 17
Research Design.............................................................................................................................. 18
Shared auditor office ................................................................................................................... 19
Lender loan loss provision validity ............................................................................................. 19
Borrower going concern reporting accuracy............................................................................... 20
Lender audit fees ......................................................................................................................... 23
Borrower audit fees ..................................................................................................................... 25
RESULTS ........................................................................................................................................... 26
Lender loan loss provision validity ................................................................................................. 27
Borrower going concern reporting accuracy................................................................................... 28
Lender Audit Fees ........................................................................................................................... 30
Borrower Audit Fees ....................................................................................................................... 31
SUPPLEMENTAL ANALYSES ....................................................................................................... 33
Cross-Sectional Analyses................................................................................................................ 33
Commercial Loan Portfolio ........................................................................................................ 33
Lender Loan Importance ............................................................................................................. 34
Alternate Measure of Lender Shared Auditor ................................................................................. 36
Continuous Measure ................................................................................................................... 36
Materiality Measure .................................................................................................................... 37
Proportion Measure ..................................................................................................................... 37
National Shared Auditor ................................................................................................................. 38
Geography Effects .......................................................................................................................... 39
Falsification Tests ........................................................................................................................... 39
3
CONCLUSION ................................................................................................................................... 40
REFERENCES ................................................................................................................................... 43
FIGURE .............................................................................................................................................. 46
APPENDIX ......................................................................................................................................... 47
TABLES ............................................................................................................................................. 49

4
ABSTRACT

Auditor competency is a key element in explaining the supply of audit quality, yet our
understanding of the drivers of auditor competency in the archival literature is limited. This
study uses an archival approach to examine whether sharing auditors among related firms results
in information spillovers that improve audit quality. I find that audit quality improves for both
borrowers and lenders who share the same auditor office. Specifically, lenders who share an
auditor office with their borrowers have more accurate loan loss provisions, especially lenders
with larger commercial loan portfolios; and borrowers who share an auditor office with their
lender are less likely to receive a clean audit opinion just prior to bankruptcy, especially
borrowers with larger loans. I also find weak evidence that lenders and borrowers who share
auditor office pay higher audit fees. Overall, these findings are consistent with shared auditors in
banking relationships developing client specific knowledge that is transferrable across clients
and industries.
Keywords: shared auditors, auditor competencies, audit quality, banking relationship

5
Do shared auditors improve audit quality? Evidence from banking relationships.

I. INTRODUCTION

Both academics and standard setters have long recognized the importance of competency
in maintaining audit quality.
1
Watts and Zimmerman (1982) argue that audit quality is a function
of both auditor incentives and their competencies. The PCAOB’s Quality Control standard
defines auditor competency as the knowledge, skills, and abilities that enable a practitioner-in-
charge to be qualified to perform an accounting, auditing, or attestation engagement; where
competencies are typically gained through their experiences in working with clients.
2
Although
auditor competency is a critical element of audit quality, insight into the factors that explain
auditor competency remain somewhat limited in the archival literature.

This study examines
whether auditing related parties involved in banking relationships improves auditor competency
by helping the auditor better understand both the lender and borrower risks.
3

I conjecture that auditing related firms in banking relationships is likely to improve
auditors’ competencies through information transfers.
4
Auditing the lender is likely to improve
the auditor’s competency in auditing the borrower because information of the lender’s loan
portfolio provides auditors with additional insights into the borrower’s business and industry,
and an external benchmark to assess the borrower’s financial condition. Similarly, auditing the
borrower is likely to improve the auditor’s competency in auditing the lender because knowledge
of the borrower’s financial condition helps the auditor evaluate the collectability of the lender’s

1
Audit quality is the assurance of financial reporting quality. Higher audit quality provides greater assurance that
the financial statements faithfully reflect the firm’s underlying economics, conditioned on its financial reporting
system and innate characteristics (DeFond & Zhang, 2014).
2
PCAOB’s Quality Control Section 40
3
The term “banking relationship” refers to the borrower and lender with an outstanding loan between them.
4
Auditor competency has many different dimensions. In this study, I focus on how information spillovers affects the
knowledge element of auditor competency.
6
loan portfolio.
5
This can improve the audit of the lender’s loan loss reserve, the largest and most
significant accrual for banks (Beatty & Liao, 2013). Information spillovers from auditing both
the borrower and lender increase auditor competency, which is expected to improve the design of
audit procedures and analysis of results, thereby resulting in higher audit quality.
6

It is not obvious, however, that auditors are able to transfer client-specific knowledge
across clients. One barrier to information transfers is that they require coordinated effort within
the auditor office. The audit team auditing the lender (borrower) must be aware of, and have
access to, the audit of the borrower (lender).
7
Another potential barrier is the AICPA’s code of
ethics. Section 301 of the AICPA Code of Professional Conduct restricts auditors from
disclosing confidential client information. This means, for example, that an auditor who becomes
aware of a large loan loss provision for a client-borrower while auditing a lender, is unable to use
that evidence to justify the issuance of a going concern opinion to the borrower.
8
On the other
hand, the U.S. Supreme Court has held auditors legally liable for damages that result from
misleading financial statements even if the statements are in strict compliance with the “letter” of
GAAP (Ball, 2009).
9
This precedent setting case creates incentives for auditors to make use of
all available information because compliance with GAAS, GAAP, or AICPA professional
standards is not sufficient to ensure legal compliance (Cashell & Fuerman, 1995). For example, a

5
I use the term “knowledge” to refer to information stored in long-term memory, where knowledge content is the
quantity and specific pieces of information (Bonner, 2008). I use the terms knowledge and knowledge content
interchangeably throughout this study.
6
Information spillovers from shared auditors reduces information asymmetry between the auditor and its clients.
This information and experience increases the auditor’s competency via knowledge.
7
In practice, access to confidential client information may be restricted to engagement team members. Auditors may
not have access to client information even within the same audit firm or audit office.
8
It is, however, acceptable for an auditor to apply the knowledge and experience developed during an audit to other
engagements as long as the details of the engagement are not disclosed (McAllister & Cripe, 2008) (ET Section
391.030). Thus, an auditor who becomes aware of a large loan loss provision for a client-borrower while
auditing a lender may use that information to design additional audit procedures for the borrower to corroborate
that information.
9
U.S. v. Simon (425 F.2d 796, 1969), United States Court of Appeals Second Circuit; Argued April 18,
1969; Decided Nov. 12, 1969, Certiorari Denied March 30, 1970.
7
jury held that client confidentiality did not supersede Arthur Andersen’s legal and ethical
obligation to report discrepancies discovered when the Denver office was auditing both parties
involved in a vending relationship.
10
In addition, the results in Dhaliwal, Lamoreaux, Litov, and
Neyland (2015) are consistent with auditors communicating with clients despite the constraints
in Section 301. Therefore, it is an empirical question whether sharing auditors improves audit
quality for both lenders and borrowers.
Banking relationships present an appealing setting to examine the effects of shared
auditors for several reasons.
11
First, this setting allows me to identify audit quality outputs that
are tightly linked to the type of knowledge auditors are likely to gain in auditing banking
relationships. Specifically, because banks are primarily interested in downside risk, knowledge
gained from auditing lenders is likely to increase the auditor’s competency in assessing the
borrower’s ability to continue as a going concern. Similarly, since auditors routinely gather
information to assess client financial health, knowledge gained from auditing borrowers is likely
to increase the auditor’s competency in auditing lenders’ loan loss provisions. This close match
between the type of information gathered during the audit and my audit quality proxies makes
for more targeted tests, and contrasts with much of the audit quality literature, which necessarily
relies on more generically-related measures of audit quality. Second, the banking relationship
setting allows for identification of the lender, borrower, and auditor in a banking relationship,
which allows for analysis of the distinct effects on audit quality for both lenders and borrowers.
This contrasts, for example, with shared auditors in supply-chain relationships, where data

10
The Fund of Funds, Limited, F.O.F. Proprietary Funds, Ltd., and IOS Growth Fund, Limited, A/K/A Transglobal
Growth Fund, Limited, Plaintiffs, v. Arthur Andersen & Co., Arthur Andersen & Co. (Switzerland), and Arthur
Andersen & Co., S.A., Defendants, No. 75 Civ. 540 (CES), United States District Court for the Southern
District of New York, 545 F. Supp. 1314; 1982 U.S. Dist. Lexis 9570; Fed. Sec. L. Rep. (Cch) P98,751, July
16, 1982.
11
I use the terms shared auditor office and shared auditor interchangeably because I focus on shared auditor offices.
I examine whether the results are driven by shared auditors (irrespective of whether they share offices) in
sensitivity analyses.
8
limitations necessarily restrict the analysis to relations with the largest customers (Johnstone, Li,
& Luo, 2014).
12
Finally, this setting provides insights about audit quality for banks. Financial
institutions are an important, but under-researched, type of firm in the audit literature.
I define shared auditors as instances in which the borrower and lender in a banking
relationship receive an audit opinion from the same auditor office.
13
I focus on shared auditor
offices because information transfers are more likely to occur at the office level. Information
spillover may occur from audit team members working on both the lender and borrower audits
(i.e. engagement quality review partner, project team specialists, manager, etc.) or through
informal communication with auditors on the related audit. Using banking relationships from
DealScan from 2000 to 2012, I compare audit quality between lenders and borrowers that have
the same auditor office (hereafter referred to as “shared auditors”) with those that have different
auditors or the same auditors from different offices. Figure 1 presents an illustration of shared
auditors in banking relationships. I examine whether the client-specific knowledge auditors
develop when auditing both lenders and borrowers improves audit quality.
I find that shared auditors are associated with lenders who have more accurate loan loss
provisions, and with borrowers who are less likely to receive clean audit opinions just prior to
bankruptcy. In addition, I find that the effect of shared auditors on audit quality is limited to
firms where knowledge content is more likely to transfer because of their importance to the
audit: lenders with larger commercial loan portfolios and borrowers with larger loans. I also find
weak evidence that audit fees increase for both the lender and the borrower, which may reflect

12
Disclosure of customer data is voluntary and may be incomplete.
13
For example, AT&T Corp and JP Morgan Chase Bank were in a banking relationship for fiscal year 2002. They
have a shared auditor since both are audited by the PricewaterhouseCoopers New York office.
9
greater audit effort.
14
These results are robust to using a propensity score matched sample and are
consistent with an improvement in audit quality for shared auditors in banking relationships. This
suggests that shared auditors develop client specific knowledge when auditing parties in banking
relationships and this knowledge is transferred across clients.
This study contributes to the literature by providing insight into how complex audit team
interactions may affect auditor competency and incentives.
15
Although auditor competency is a
key input in the supply of audit quality, DeFond and Zhang (2014) point out that the actual
elements of auditor competency are currently under-researched in the archival literature. Prior
research has focused on office size (i.e. Big N) and industry specialization as the primary
elements of auditor competency. However, it is not clear why Big N audit firms are higher
quality and the industry specialization literature makes strong assumptions about the mechanisms
that improves audit quality. This study uses a unique setting to examine client specific
knowledge developed in auditing banking relationships as an alternate source of auditor
competency. Knowledge in related firms may improve the integration of audit evidence that in
turn enhances the auditor’s understanding of the clients’ overall business. This knowledge
improves audit quality for both parties in the banking relationship and is incremental to industry
expertise.
Second, this study contributes to the literature by showing that auditor knowledge is
transferrable across client firms and industries. Auditor expertise is typically presumed to be

14
Bell, Landsman, and Shackleford (2001) find that auditors adjust the number of hours and not the fee per hour to
reflect risk identified. This suggests that audit fees reflect audit effort and efficiency rather than a risk premium.
15
Auditor competencies are not independent of their incentives (DeFond & Zhang, 2014). Greater incentives to
supply high audit quality motivate auditors to develop competencies. Similarly, greater competency in
delivering high quality audits increases the auditor’s reputation capital, thereby providing greater
incentives to supply high audit quality.
10
industry specific.
16
However, the transferability of industry expertise across firms may be limited
since direct competitors may not want to employ the same auditor (Aobdia, 2015).
17
The findings
from this study suggests that client specific knowledge developed while auditing lenders
(borrowers) is transferred to borrowers (lenders) even though they are in different industries.
This suggests that client specific knowledge may be transferred to other audit tasks across both
clients and industries.
Another contribution of this study is that it adds to the relatively small literature on
shared auditors by showing that both clients in a shared auditor relationship can benefit.
Johnstone et al. (2014) examines shared auditors in supply-chain relationships and find that
shared auditors improve audit quality and decrease audit fees for supplier companies. Due to data
limitations, their proxy for shared auditors is a general portfolio measure and their analyses are
restricted to suppliers. More recently, Dhaliwal et al. (2015) examines the effects of shared
auditors on both the bidding firm and target firm in an acquisition and find that shared auditors
benefit the bidder at the expense of the target. Although Dhaliwal et al. (2015) examine the
reciprocal effects of shared auditors, they focus on the effects of shared auditors on acquisition
outcomes. This study extends this line of research by examining how shared auditors affect both
parties in a setting where information transfers about default and loan risks are particularly
salient, and finds that audit quality increases for both parties in the shared relationship. Finally,
the results suggest that while Section 301 proscribes sharing information across clients, I find
that such sharing can actually improve audit quality.

16
Generally accepted auditing standards (GAAS) AU-C Section 620 defines expertise as skills, knowledge, and
experience in a particular field. Bonner (2008) defines expertise as high-quality judgment and decision making.
17
For example, the AICPA Code of Professional Conduct Section 391-6.011 discusses a case where a firm objects
to being audited because “information gathered from the books and records of his or her client could be
inadvertently conveyed to competitors by employees of the CPA firm doing the audit”.
11
The findings in this study may also inform regulators and practitioners about the
mechanisms that develop auditor competency. Audit firms typically structure practice groups by
industries so that auditors can gain competencies by working on industry related engagements.
This study suggests that auditors also develop key competencies by working on related
engagements outside of an industry. The competencies developed in related engagements can
improve auditor competencies in both fields and is incremental to the competencies gained from
industry experience. Thus, while I examine shared auditors in banking relationships, the
implications of the findings are not limited to this setting.
18

The remainder of this paper is organized as follows: Section II briefly reviews the
relevant literature and develops the hypotheses. Section III introduces the sample and research
design. Section IV discusses the empirical results while Section V provides supplemental
analyses. Finally, Section VI reviews the findings and concludes.

II. LITERATURE REVIEW AND HYPOTHESIS DEVELOPMENT

Background and prior literature
Auditor competencies
There is a large stream of experimental research that explores the determinants of auditor
competency and its relation to audit quality (Bonner, 2008). In contrast, archival research has
primarily focused on relatively crude measures of auditor competency such as auditor size (i.e.
BigN) and industry specialization (industry market share leader) (Balsam, Krishnan, & Yang,
2003; J. R. Francis, Reichelt, & Wang, 2005; Krishnan, 2003; Reichelt & Wang, 2010) and
generally finds support for its effect on audit fees and audit quality (DeFond and Zhang, 2014).

18
In private conversation, a Big N auditor with many years of experience indicated that the results in this study are
consistent with his experiences as a shared auditor in construction-contractor relationships.
12
However, there is limited archival research exploring the specific mechanisms that build
competency. One potential mechanism is through increased knowledge. Consistent with
knowledge improving auditor competency, prior experimental research finds knowledge content
is positively associated with audit quality (Bonner, 2008). In this study, I use shared auditors as a
proxy for auditor knowledge in an archival setting while controlling for auditor size and industry
specialization. Shared auditors may develop knowledge content or competency when auditing
both the lender and borrower because audits of the related parties should provide more specific
and relevant knowledge, help build other required knowledge, and improve the integration of
such knowledge, which then improves audit quality. The findings in this study can complement
the experimental literature to show how shared auditors affects actual audit quality outcomes.
Shared Auditor
Johnstone, Li, and Luo (2014) examine the effects of shared auditors in supply-chain
relationships on supplier’s audit quality. They measure an auditor’s “supply chain expertise” as
the ratio of shared supplier-customer auditors divided by the total number of suppliers audited by
an audit firm, which captures shared auditors within the auditor’s portfolio. They find that supply
chain expertise increases audit quality (by decreasing absolute discretionary accruals and
lowering the likelihood of restatements or managing earnings to meet or beat analysts’ forecasts)
and lowers audit fees for supplier companies. These effects are driven by supply chain expertise
in the revenue cycle. In comparison, I measure audit quality at the individual client (both
borrowers and lenders) level, and use audit quality measures that are more directly linked to
auditor’s knowledge spillover from auditing both lenders and borrowers (lenders’ loan loss
provision, borrowers’ going concern opinions, and audit fees).
13
Dhaliwal, Lamoreaux, Litov, and Neyland (2015) examines the effects of shared auditors
in mergers and acquisitions. In this setting, shared auditors facilitate the flow of information
between bidders and targets, which mitigates information asymmetry problems. They find that
shared auditors in mergers and acquisitions primarily benefit the bidder with higher bidder
returns and completion rates. These benefits come at the expense of the target, which earns lower
event returns and lower deal premiums. While their findings are consistent with knowledge
spillover from auditing related firms, Dhaliwal et al. (2015) focus on the effects of shared
auditors and acquisition outcomes. In contrast, this study examines the implication of knowledge
spillover on audit quality.
Hypothesis Development
I examine whether information spillover in related engagements is a mechanism for
improving audit quality. Industry experience is typically presumed to be a source of auditor
competencies because auditors have greater knowledge of industry business and accounting
practices (Dopuch & Simunic, 1982). I extend this line of research by studying whether working
in related engagements in different industries can be an alternate source of information and
knowledge. Specifically, I explore the potential transfer of knowledge when engagements are
related through banking relationships.
Auditing the borrower in a lender’s loan portfolio may provide the auditor a better
understanding of the borrowers’ financial condition and credit risk, which should improve the
auditor’s assessment of the collectability of the loans and whether loan loss provisions are
necessary. Loan loss provisions are typically a bank’s largest operating accrual (Altamuro &
Beatty, 2010; Beatty & Liao, 2013; GAO, 1991, 1994) and is the accrual most subject to
discretion. Audit procedures to review loan credit include consideration of the following factors:
14
context of the loan, payment history, financial condition, past and expected future cash flows,
pledged collateral, and guarantors (Thibodeau, 2003). According to FASB ASC 310-10-S99-4
(formerly Staff Accounting Bulletin 102), the method used to estimate loan losses should include
procedures to reduce differences between estimated losses and actual subsequent charge-offs. As
such, I measure the audit quality of lenders by examining the validity or accuracy of loan loss
provisions, which is measured using the association between loan-loss provisions and actual
charge-offs (hereafter referred to as “loan loss provision validity” or “loan loss provision
accuracy”). I expect shared auditors to improve the lenders’ audit quality through a stronger
association between current loan-loss provisions and subsequent actual charge-offs.
H1: Shared auditors are positively associated with lender loan loss provision validity.
Auditing the lender may provide the auditor a better understanding of the borrower’s
credit risk. Banking relationships provide an interesting setting for this study because banks are
in the business of assessing downside risk and may have a better understanding of the borrower’s
credit worthiness, credit risk, and default risk. The necessity of a loan loss provision suggests
that there may be doubt about the borrower’s ability to continue as a going concern. As a result,
the auditor may gain additional insight on the financial condition of the borrower when they are
auditing the lender’s loan loss provision. A shared auditor may also have additional insight on
actual or potential covenant violations and the borrower’s ability to secure (additional) financing.
Since shared auditors may have more insight into borrower risks related to the loan,
shared auditor going-concern reporting may be more accurate than going-concern reporting by
different auditors. Auditors are responsible for issuing a going concern when there is substantial
doubt about a firm’s ability to continue as a going concern for the next 12 months. Investor
reaction to going concern audit reports indicates that audit reports are more informative when the
15
reports disclose problems with securing financing or violations of debt covenants (Menon &
Williams, 2010). Auditors routinely make Type II errors (i.e., issuance of a clean opinion in the
year prior to bankruptcy) about 50% of the time (Carson et al., 2012; Raghunandan & Rama,
1995) and Type I errors (i.e., issuance of a GC opinion in the absence of bankruptcy within the
subsequent year) about 90% of the time (Geiger, Raghunandan, & Rama, 2005). Carson et al.
(2012) suggest that Type I errors are due to the difficulty in determining the subsequent viability
of firms that receive a going concern; the high Type II error rate is due to auditors lack of
expertise or specialized knowledge to assess bankruptcy risk (Arnold, Collier, Leech, & Sutton,
2001; Carson et al., 2012). If information from the lender helps the auditor assess downside risk,
I expect greater accuracy in the issuance of going concerns for borrowers; shared auditors are
more likely to reduce Type II errors since these errors result from lack of knowledge.
H2: Shared auditors are positively associated with going concern reporting accuracy.
In addition to evaluating the effects of shared auditors on loan loss provision and going
concern opinions (output measures), I also examine whether shared auditors affect audit fees (an
input measure) to infer audit quality. I use audit fees as a measure to gather insight into the
supply of audit quality (by the auditor) and the demand for audit quality (by audit committees)
since clients must choose audit quality based on observable inputs. However, audit fees are a
relatively noisy measure of audit quality because inputs may not directly translate into outputs;
fees also capture a combination of audit effort, risk premia, and audit efficiency (DeFond &
Zhang, 2014).
Shared auditors will be positively associated with audit fees if it increases the effort, and
therefore quality, of the audit. From the supply side, shared auditors increase effort when
auditing parties in a banking relationship because they have more information to identify
16
potential risk factors. This translates into additional audit effort since the auditor will need to
perform procedures to address the risks identified, which should be associated with higher audit
fees and audit quality. Without this information, auditors may overlook potential business risk,
which would result in lower audit effort and fees. From the demand side, lender audit
committees recognize that shared auditors have a better understanding of the business and serve
as better monitors. Borrower firms, on the other hand, are willing to pay a premium to signal
high quality financial statements to obtain financing. These factors suggest that shared auditors
should be positively associated with audit fees for both the lender and the borrower.
On the other hand, shared auditors will be negatively associated with audit fees if it
increases the efficiency of the audit. Common elements between auditing the lender’s loan loss
reserves and the borrower’s going concern assessment may reduce the audit procedures
necessary to verify certain elements of the financial statements. However, the AICPA’s
Professional Code of Conduct restricts auditors from directly transferring audit evidence
obtained from one engagement to another engagement. Although eliminating unnecessary
procedures may increase the efficiency of the audit, efficiencies are limited because the audit
team must gather sufficient audit evidence to support their decision without referring to the
related engagement. Also, this would only result in lower audit fees if the auditor chooses to pass
on the savings to the client.
Overall, shared auditors may increase audit fees because of additional effort required to
address downside risks. On the other hand, shared auditors can decrease audit fees because of the
efficiency in performing both audits. Although there are arguments for both a positive and
negative association between shared auditors and audit fees, I believe the arguments for a
positive association are stronger. Knowledge spillover from banking relationships likely relates
17
to the identification of additional downside risk, which typically requires additional procedures.
Since the AICPA Professional Code of Conduct restricts auditors from transferring audit files
between engagements, this knowledge likely to require additional effort. Further, any efficiencies
from the information spillover would decrease audit fees only if the auditor chooses to share the
discount with the client. As such, I hypothesize that shared auditors are positively associated
with audit fees for the lender and borrower.
H3: Shared auditors are positively associated with lender audit fees.
H4: Shared auditors are positively associated with borrower audit fees.

III. RESEARCH DESIGN AND SAMPLE

Sample
Data on banking relationships from 2000 through 2012 was obtained from DealScan. I
retain all sole lending relationships and the lead bank of all syndicate loans, which results in
21,824 banking relationships.
19
Since lenders have banking relationships with multiple
borrowers, I use different samples for lenders and borrowers. Borrower financial data was
obtained from Compustat and CRSP and auditor information was obtained from Audit Analytics,
which resulted in 15,037 banking relationships or 78,540 borrower firm-years and 1,475 lender
firm-years.
20
I excluded borrowers incorporated outside of the United States or Canada,
American depositary receipt firms, financial and utility firms, and firms with non-BigN auditors,
resulting in 9,574 banking relationships or 50,879 borrower firm-years and 869 lender firm-
years. Lender financial information was obtained from the Federal Reserve Bank of Chicago FR-
Y9C reports resulting in 7,692 banking relationships or 43,193 borrower firm-years and 469

19
A lender is designated a sole lender or lead bank if its lender role is one of the following: Admin Agent, Agent,
Arranger, Lead bank, or Sole lender based on Bharath, Dahiya, Saunders, and Srinivasan (2009).
20
Dealscan-Compustat link data was provided by Chava and Roberts (2008).
18
lender firm-years.
21
After obtaining data for all relevant control variables, there are 406 firm-
years for the lender audit quality sample, 26,130 firm-years for the borrower audit quality
sample, 389 firm-years for the lender audit fee sample, and 39,415 firm-years for the borrower
audit fee sample. Table 1, Panel A presents the details of the sample selection.
Table 2, Panel B provides details on the sample of firm-years with shared auditors.
Shared auditors occur in 1,184 firm-years. Shared auditor firm years approximate 10% of the
shared sample in the early part of the decade from 2000 through 2003. From 2004 through 2012,
shared auditors approximate 7% of the shared auditor sample. Changes in shared auditor offices
is attributable to a number of factors including lender switches to another BigN or BigN office,
borrower switches to another BigN or BigN office, and the termination of loans. Other factors
include firm switches to a non-BigN auditor or termination of coverage on Compustat or Audit
Analytics because the firm went private, which coincides with the enactment of Sarbanes-Oxley
Act (SOX). This is consistent with the evidence in the Leuz, Triantis, and Wang (2008) which
documented a spike in firms that went dark due to poor prospects, distress, and compliance costs
after SOX. It may also reflects the auditors’ rebalancing of their client portfolios as a result of the
Andersen indictment (Landsman, Nelson, & Rountree, 2009).
Research Design
I examine the effects of having a shared auditor on several output and input audit quality
variables. Output measures provide insight into the level of audit quality actually delivered while
input measures provides an alternative way to infer audit quality because audit clients must
choose audit quality based on observable inputs (DeFond & Zhang, 2014). The output audit
quality variables are loan loss provision validity for the lender and going concern reporting

21
Federal Reserve Bank-CRSP link data was provided by the Federal Reserve Bank of New York (2014).
19
accuracy for the borrower. The input audit quality variable is audit fees for both the lender and
the borrower. Appendix A presents a summary of all variables used in this study.
Shared auditor office
Shared auditor office is defined separately for the borrower and the lender. Since the
borrower sample tracks relationships on an individual basis, SHAREAUDOFF is an indicator
variable that equals one if the lender and borrower received audit opinions from the same auditor
and city, and zero otherwise. Lenders, on the other hand, may have multiple borrowers. The
number of shared auditor relationships per lender ranges from one to 33 with a mean (median) of
6 (3). The lender sample includes 202 firm-years with at least one shared auditor relationship.
However, it is worth noting that a single shared auditor relationship may not be meaningful to
the lender. To capture meaningful shared auditor relationships in the lender sample,
LSHAREAUDOFF is an indicator variable that equals one if the lender has greater than the
median number of shared auditor relationships, and zero otherwise.
22

Lender loan loss provision validity
I test the effect of shared auditors on lender audit quality by examining the relation
between current period loan loss provisions and next period net charge offs. Following Altamuro
and Beatty (2010), I estimate the following model (firm index omitted for brevity):
CHGOFF
t+1
= α + β
1
LSHAREAUDOFF
t
+ β
2
LSHAREAUDOFF
t
*LLP
t
+
β
3
LSPECIALIST
t
+ β
4
LSPECIALIST
t
*LLP
t
+ β
5
LLP
t
+ β
6
LLSIZE
t
+
β
7
NONACC
t
+ ϵ (1)
where CHGOFF is the next period’s net loan charge-offs scaled by total assets. I use a one-year
horizon for measuring future losses because the Office of the Comptroller and many banks

22
I use a binary variable to define a shared auditor in both the lender and borrower analyses for consistency. I use
several alternate measures of lender shared auditor in supplemental analyses.
20
consider one year coverage of losses an adequate reserve for most pools of loans (Office of the
Comptroller of the Currency, 1998). I use net charge-offs because banks can obscure income
through gross charge-offs or recoveries (Liu & Ryan, 2006). LSHAREAUDOFF is an indicator
variable that equals one if the lender has greater than the median number of shared auditor
relationships and zero otherwise, LLP is the loan loss provision during year t scaled by beginning
total assets, and LSPECIALIST is an indicator variable that equals one if the lender’s auditor has
the largest market share of audit fees in a two-digit SIC category year, and zero otherwise. The
primary coefficient of interest is β
2
. A positive coefficient indicates that shared auditors increase
the association between current loan loss provisions and actual charge-offs. I also include the
LSPECIALIST variable to control for industry specialist expertise. Following prior research
(Altamuro & Beatty, 2010), I include firm size and non-performing loans as control variables.
LLSIZE is the log of total assets at the beginning of year t, and NONACC is non-performing
loans at the end of year t scaled by beginning total assets. Non-performing loans are loans that
are less than 90 days past due and not accruing interest.
Borrower going concern reporting accuracy
I test the effect of shared auditors on the borrower’s audit quality by examining the
auditor’s accuracy of going concern reporting. There are two types of going concern reporting
misclassifications: Type I misclassification occurs when the auditor issues a going concern to a
firm that does not subsequently fail; Type II misclassification occurs when the auditor does not
issue a going concern and the client fails. To assess the accuracy of going concern reporting, I
examine the propensity of issuing a going concern on two different samples: 1) firms that have
not subsequently filed for bankruptcy as of December 2013 to examine Type I misclassification
and 2) firms that filed for bankruptcy to examine Type II misclassification. An observation is
21
included in the bankrupt sample if the firm files for Chapter 7 or Chapter 11 bankruptcy within
one year of the fiscal year end or audit opinion signature date according to Audit Analytics. I use
the following logit model (firm and year index omitted for brevity) to examine the propensity of
issuing a going concern:
OPINION = α + β
1
SHAREAUDOFF + β
2
SPECIALIST + β
3
LSPECIALIST +
β
4
ZMIJEVSKI + β
5
BLLOSS + β
6
AGE + β
7
TOTLEV + β
8
CLEV + β
9
CFO +
β
10
RISKY + β
11
SIZE + β
12
ROA + β
13
MB + β
14
SHAREISSUE +
β
15
FUTUREFINANCE + β
16
INVESTMENTS + β
17
RLAG + ϵ (2)
where OPINION is an indicator variable equal to one the auditor issued a going concern opinion,
and zero otherwise. SHAREAUDOFF is an indicator variable equal to one if the lender and
borrower received audit opinions from the same auditor office, and zero otherwise. β
1
is the
primary coefficient of interest. A negative coefficient indicates that shared auditors decrease
going concern reporting while a positive coefficient indicates that shared auditors increase going
concern reporting. If shared auditors are more accurate at issuing going concern opinions, β
1

should be negative for firms that have not filed for bankruptcy and positive for firms that
subsequently file for bankruptcy. SPECIALIST
t
is an indicator variable that equals one if the
borrower’s auditor has the largest market share of audit fees in a two-digit SIC category year,
and zero otherwise and LSPECIALIST is an indicator variable that equals one if the lender’s
auditor has the largest market share of audit fees in a two-digit SIC category year, and zero
otherwise. The SPECIALIST and LSPECIALIST variables are included to control for industry
specialist expertise. Contrary factors indicating financial distress and mitigating factors that
would mitigate against a going concern report are included as control variables based on prior
research (DeFond, Raghunandan, & Subramanyam, 2002; Geiger et al., 2005; Raghunandan &
22
Rama, 1995; Reynolds & Francis, 2000). Financial distress is captured using several different
variables. ZMIJEWSKI is a bankruptcy measure with higher values indicating higher probability
of bankruptcy (Zmijewski, 1984). BLLOSS is an indicator variable that equals one if the firm had
a loss in the prior year, and zero otherwise. This variable is included because firms that
experience losses in multiple periods are more likely distressed (Reynolds & Francis, 2000).
AGE is the log of number of years the company has been publicly traded and is included because
younger firms are more prone to failure. TOTLEV is total liabilities over total assets at the end of
the fiscal year. CLEV is the change in leverage during the year. These variables are included to
proxy for closeness to debt covenant violation (DeFond et al., 2002), which should increase with
a going concern assessment. However, leverage may also increase when a firm secures
financing, which could decrease the likelihood of getting a concern. CFO is operating cash flow
scaled by total assets. This variable is included because changes in liquidity affects financial
distress. RISKY is an indicator variable equal to one if the company operated in a risky
industry.
23
This variable is included because high tech industries may have higher bankruptcy
risk (Raghunandan & Rama, 1995). However, I do not predict a sign because of inconclusive
evidence in prior studies.
The following mitigating factors are also included as control variables. SIZE is the log of
total assets and is included because large firms are more likely to avoid bankruptcy (Reynolds &
Francis, 2000). ROA is net income scaled by average total assets. MB is the market value of
equity scaled by book value of equity at the end of year t. These variables are included to capture
firm performance and growth opportunities and should be negatively related to bankruptcy.
SHAREISSUE is an indicator variable equal to one when a firm issued equity in the fiscal year,
and zero otherwise. FUTUREFINANCE is an indicator variable equal to one when the firm

23
Risky industries are firms with SIC codes 2833, 2836, 3570, 3577, 3600, 3674, 7372, 7379, 8731, and 8734.
23
issues equity or long-term debt in the subsequent year, and zero otherwise. These variables are
included because financing, refinancing, and increase in ownership equity are managerial actions
that mitigates bankruptcy (Reynolds & Francis, 2000). INVESTMENTS are short- and long-term
investment securities (including cash and cash equivalents) scaled by total assets at year-end and
is included as an ex ante measure of plans to sell assets (DeFond et al., 2002). Although proceeds
from asset sales may be a mitigating factor for firms operating as a going concern, asset sales
may be a contrary factor for firms experiencing extreme financial distress. Asset sales for these
firms may be due to pressure to sell operating assets to meet its maturing obligations (Venuti,
2004) which result in “fire sales” of assets at heavily discounted prices. RLAG is the number of
days between the fiscal year-end and the opinion date and is included because going concerns are
associated with longer reporting delays (Raghunandan & Rama, 1995).
Lender audit fees
The effects of shared auditors on audit fees were estimated using the following audit fee
model specific to banks (firm and year index omitted for brevity):
LAUDITFEE = α + β
1
LSHAREAUDOFF + β
2
LSPECIALIST + β
3
LSIZE + β
4
LLOSS +
β
5
SECURITIES + β
6
NONPERFORM + β
7
CHGOFFCY +
β
8
COMMLOAN + β
9
CONSUMERLOAN+ β
10
RELOAN + β
11
INTANG +
β
12
CAPRATIO + β
13
EXEMPT+ β
14
SENSITIVE + ϵ (3)
where LAUDITFEE is the natural log of lender audit fees and LSHAREAUDOFF is an indicator
variable that equals one if the lender has greater than the median number of shared auditor
relationships. LSPECIALIST is an indicator variable that equals one if the lender’s auditor has
the largest market share of audit fees in a two-digit SIC category year, and zero otherwise. The
specialist variable is included to control for industry specialist expertise. β
1
is the primary
24
coefficients of interest. A significant coefficient indicates that shared auditors are associated with
lender audit fees. Control variables include measures for firm size, complexity, and risk
(liquidity, capital, and credit) and are based on banking firms audit fee models (Fields, Fraser, &
Wilkins, 2004; Kanagaretnam, Krishnan, & Lobo, 2010). LSIZE is the natural log of total assets.
This is a measure of lender size and should be positively associated with fees. LLOSS is an
indicator variable that equals 1 if net income is negative, and zero otherwise. SECURITIES is
defined as one less total securities deflated by total assets. Securities are liquid assets and is
included as a measure of liquidity risk. The following variables are measures of bank credit risk.
NONPERFORM is non-performing loans divided by lagged total loans. CHGOFFCY is net
charge-offs deflated by beginning total assets. COMMLOAN is total commercial and agricultural
loans divided by total loans. CONSUMERLOAN is total consumer loans divided by total loans.
RELOAN is total real estate loans divided by total loans. Audit fees typically increase with
measures of risk. However, the association of each loan type varies depending on how the
lender’s portfolio reflects overall risks. Hence, there is no prediction for COMMLOAN,
CONSUMERLOAN, or RELOAN. I include INTANG and CAPRATIO as measures of capital risk.
INTANG is intangible assets divided by total assets. INTANG is a measure of complexity and
risk-taking. It also captures banks with acquisition activity, which requires greater audit effort
and higher capital risk since goodwill decreases regulatory capital. This variable increases audit
fees. CAPRATIO is total risk-adjusted capital ratio. Higher values of CAPRATIO can increase
regulatory pressure, which can in turn increase audit fees. EXEMPT is an indicator variable that
equals one if a bank is exempt from FDICIA and Section 404 of SOX, and zero otherwise.
Exempted banks tend to pay lower audit fees. SENSITIVE is rate-sensitive assets minus rate-
25
sensitive liabilities. This variable is included to control for rising interest rates during 2000,
which benefited asset-sensitive banks and should decrease audit fees.
Borrower audit fees
The effects of shared auditors on borrower audit fees were estimated using the following
audit fee model:
BAUDITFEE = α + β
1
SHAREAUDOFF + β
2
SPECIALIST + β
3
LSPECIALIST+ β
4
SIZE +
β
5
CURR + β
6
QUICK + β
7
ROA + β
8
LEV + β
9
BLOSS +
β
10
NUMOPESEGMENTS + β
11
NUMFOREIGNSEGMENTS +
β
12
AUDITORSWITCH + β
13
OPINION + β
14
DECYE + ϵ (4)
where BAUDITFEE is the natural log of borrower audit fees and SHAREAUDOFF is an
indicator variable equal to one if the lender and borrower received audit opinions from the same
auditor office, and zero otherwise. β
1
is the primary coefficient of interest. A significant
coefficient indicates that shared auditors are associated with lender audit fees. SPECIALIST is an
indicator variable that equals one if the borrower’s auditor has the largest market share of audit
fees in a two-digit SIC category year, and zero otherwise and LSPECIALIST is an indicator
variable that equals one if the lender’s auditor has the largest market share of audit fees in a two-
digit SIC category year, and zero otherwise. The specialist variables are included to control for
industry specialist expertise. Control variables are included based on the extant literature on the
determinants of audit fees. SIZE is the natural log of assets and should increase with larger
clients. CURR is the ratio of current assets to current liabilities and QUICK is the ratio of current
assets excluding inventories to current liabilities. These are audit risk variables and should be
negatively related to audit fees since firms with higher ratios are less risky. ROA is net income
scaled by average total assets. This should be negatively related to audit fees since more
26
profitable clients are less risky for auditors. LEV is the sum of current and long-term debt scaled
by average total assets and should be positively related to audit fees. BLOSS is an indicator
variable equal to one if the firm has negative net income in the current year, and zero otherwise.
Loss firms pose greater audit risk and should be positively related to audit fees.
NUMOPESEGMENTS is the natural logarithm of one plus the number of operating segments and
NUMFOREIGNSEGMENTS is the natural logarithm of one plus the number of foreign segments.
These variables capture the complexity of the client and should be positively related to audit
fees. AUDITORSWITCH is an indicator variable equal to one if the firm switched auditors for the
current year, and zero otherwise. Auditor switches can either have a positive or negative effect
on audit fees. On one hand, auditor switches are positively associated with audit fees because it
is an initial year audit and the audit firm incurs higher costs because first year audit engagements
often require more time and effort than recurring audit engagements. On the other hand, audit
fees may be negatively associated with audit fees because of low-balling. OPINION is an
indicator variable equal to one if the firm received a going concern opinion, and zero otherwise.
This variable should be positively associated with audit fees because a going concern assessment
entails additional risk and procedures. DECYE is an indicator variable equal to one if the firm
has a December fiscal year-end, and zero otherwise. This should be positively related to audit
fees because clients will be charged a premium for work performed during busy season.

IV. RESULTS

I examine the effects of having a shared auditor on loan loss provision validity for the
lender, going concern reporting accuracy for the borrower, and audit fees for both the lender and
27
the borrower. These reflect both input and output audit quality measures. Univariate and
multivariate analyses are presented by each type of test.
Lender loan loss provision validity
Table 2, Panel A provides descriptive statistics on the subsample of firms with and
without a shared auditor office. The mean (median) size lender with shared auditors is 18.8
(18.8) billion dollars in total assets and is significantly larger than lenders without shared
auditors with 17.1 (17.3) billion in total assets. Mean (median) charge-offs are 0.01 (0.008)
billion for shared auditors are larger than charge-offs of 0.008 (0.004) billion for firms without
shared auditors. The loan loss provisions are 0.006 (0.004) billion for lenders with shared
auditors and 0.006 (0.003) billion for lenders without shared auditors. The number of shared
auditor relationships per lender ranges from zero to 33 with a mean (median) of 6 (3).
There is a potential self-selection bias since each firm has the choice to hire a shared
auditor. Since the treatment and control groups are not random, I conduct analyses using both the
full sample and a propensity-score matched sample to address endogeneity concerns of auditor
selection. I use propensity score matching with caliper matching and included all of the second
stage variables in the first stage regression. Table 2, Panel B shows the covariate balance before
and after matching. I verified that covariates are balanced across the treatment and control group.
Results for the propensity score matched sample are qualitatively similar to those of the full
sample and are presented in each table along with the full sample results.
Table 2, Panel C provides the results of the loan loss model. The primary coefficient of
interest is β
2
or the interaction between LSHAREAUDOFF and LLP in the full (matched) sample.
There is a significantly positive association (p-value < 0.05) between this variable and loan
charge-offs in both samples. This indicates that shared auditors have a stronger association
28
between accrual and operating activity relative to different auditors or offices. The association
between LSPECIALIST*LLP and loan loss provisions and charge-offs are positive and weakly
significant (insignificant) in the full (matched) sample. Loan loss provisions, lender size, and
non-performing loans are also positively associated with charge-offs. These results are consistent
with the prediction that shared auditors are positively associated with the validity of loan loss
provisions.
Borrower going concern reporting accuracy
In tests for borrower accounting quality, I conduct analyses using the full sample and a
matched distressed sample. Following prior literature, the distressed sample is composed of
financially distressed firms that have current year losses or negative operating cash flows
(DeFond et al., 2002) resulting in a sample of 6,105 borrower firm years. Multivariate analyses
are conducted using two samples of firms that were created based on whether or not the firm
filed for bankruptcy in the subsequent year.
Table 3, Panel A presents descriptive statistics on the subsample of firms with and
without a shared auditor office. The mean (median) size borrower with shared auditors is 7.8
(7.6) and is significantly larger than borrowers without shared auditors size 7.4 (7.4). Mean
going concern opinions are 1% for borrowers without shared auditors which is significantly
higher than 0.7% for borrowers with shared auditors.
Table 3, Panel B presents the results of the going concern model with partitions on
whether or not the firm subsequently filed for bankruptcy. Shared auditors are directionally
consistent with issuing more accurate going concern opinions; SHAREAUDOFF is significantly
(p < 0.05) positively associated with issuing a going concern opinion for firms that subsequently
filed for bankruptcy and is negatively associated with issuing a going concern for firms that have
29
not filed for bankruptcy. Since knowledge transfers from auditing banking relationships provides
information about downside risk, shared auditors are more likely to reduce Type II errors (i.e.
issuing a clean opinion in the year prior to bankruptcy). Shared auditors should also reduce
Type I errors that are related to financing; however, the effect of shared auditors may be
insignificant because Type I errors can also arise from the auditor’s evaluation of internal matters
(i.e. loss of key management or personnel, substantial dependence on the success of a particular
project, labor difficulties, etc.) or other external events (i.e. legal proceedings, legislation, loss of
a key customer or supplier, natural disaster, etc.). Although insignificant, the negative
association with Type I errors provides evidence that the increase in the accuracy of going
concern reporting for firms that subsequently file for bankruptcy is not necessarily due to the
issuance of more going concern opinions. SPECIALIST is significantly (p < 0.01) negatively
associated with going concerns for firms that have not filed for bankruptcy. LSPECIALIST are
not significantly associated with going concern reporting for either group of firms.
Most of the control variables are fairly consistent with predictions and are the same for
both bankrupt and non-failing firms. One interesting difference is the coefficient on
INVESTMENTS, which is significantly (p < 0.01) positive for bankrupt firms and significantly (p
< 0.01) negative for non-failing firms. The negative coefficient for non-failing firms suggests
that this variable captures firm plans to sell assets and provide evidence of management’s plans
to mitigate going concern risk. The positive coefficient on bankrupt firms, however, suggests that
this variable is capturing potential fire sales for firms under extreme distress. Fire sales occur in
extreme conditions when borrowers and lenders are unable to negotiate or renegotiate contracts
because additional cash flows cannot be pledged and a high-valuation buyer is unavailable
(Shleifer & Vishny, 2010).
30
Using the results of the distressed sample, the probability of a going concern for a shared
auditor is 0.876 and without a shared auditor is 0.468. The marginal effect of shared auditor
holding all covariates at the mean is 0.408. Overall, the results are consistent with the prediction
that shared auditors are associated with the accuracy of going concern reporting for bankrupt
firms. This implies that shared auditors are less likely to make Type II errors (i.e. issuing a clean
opinion in the year prior to bankruptcy). Since Type II errors result from lack of knowledge or
expertise (Arnold et al., 2001; Carson et al., 2012), this finding is consistent with the prediction
that shared auditors have information or knowledge from auditing banking relationships that
improves their going concern assessments for firms that subsequently file for bankruptcy.
Lender Audit Fees
Table 4 presents the results for lender audit fees. Table 4, Panel A provides descriptive
statistics on the subsample of firms with and without a shared auditor office. The mean (median)
size lender with shared auditors is 18.9 (18.9) and is significantly larger than lenders without
shared auditors size 17.2 (17.3). This corresponds with mean (median) lender audit fees of 15.8
(15.6) for lenders with shared auditors and 14.4 (14.4) for lenders without shared auditors.
In tests for lender audit fees, I conduct analyses using the full sample and a propensity-
score matched sample to address endogeneity concerns of auditor selection. I use propensity
score matching with replacement and caliper matching and include all of the second stage
variables in the first stage regression. Table 4, Panel B shows the covariate balance before and
after matching. Covariate balance across the treatment and control group improves after
matching. Results for the propensity score matched sample are qualitatively similar to those of
the full sample and are presented in each table along with the full sample results.
31
Table 4, Panel C presents the results of the lender audit fee model, which is a fee model
developed specifically for banks. The results in column (1) and (2) indicate that the shared
auditor variable is positive and significantly (p < 0.05) associated with lender audit fees in both
the full and propensity score matched sample. This suggests that shared auditors charge a fee
premium presumably for more effort, and the lender’s audit committee is willing to pay this
premium because shared auditors provide better monitoring.
Consistent with the prior literature (Craswell, Francis, & Taylor, 1995; Ferguson, Francis,
& Stokes, 2003; J. R. Francis et al., 2005), specialists are associated with a fee premium. Audit
fees are positively associated with size, losses, commercial loans, intangibles, and the capital
ratio and negatively associated with banks exempted from the Federal Deposit Insurance
Corporation Improvement Act to evaluate internal control over financial reporting. Increases in
consumer loans and real estate loans are negatively associated with audit fees. Since loan
portfolios are comprised of commercial loans, consumer loans, and real estate loans, increases in
these types of loans may be less risky than increasing the portfolio of commercial loans. These
results are consistent with the prediction that shared auditors are positively associated with
lender audit fees.
Borrower Audit Fees
Table 5 presents the results for borrower audit fees. Table 5, Panel A presents descriptive
statistics on the subsample of firms with and without a shared auditor office. The mean (median)
size borrower with shared auditors is 7.8 (7.5) and is significantly larger than borrowers without
shared auditors size 7.5 (7.5). This corresponds with mean (median) borrower audit fees of 14.5
(14.5) for borrowers with shared auditors and 14.3 (14.3) for borrowers without shared auditors.
32
In tests for borrower audit fees, I conduct analyses using the full sample and a propensity-
score matched sample to address endogeneity concerns of auditor selection. I include lender size
and borrower size along with all of the second stage variables in the first stage regression. I use
propensity score matching with replacement and caliper matching. Table 5, Panel B shows the
covariate balance before and after matching. I verified that covariates are balanced across the
treatment and control group in the weighted sample. Results for the propensity score matched
sample are qualitatively similar to those of the full sample and are presented in each table along
with the full sample results
Table 5, Panel C presents the results of the borrower audit fee model. SHAREAUDOFF is
significantly (p < 0.10) positively associated with borrower audit fees in the full and propensity-
score matched samples. Results indicate a 7.3% (6.5%), or $119k ($106k), premium for shared
auditors in the full (matched) sample, holding all other factors constant. Successfully obtaining
financing is a critical part of demonstrating the ability to continue as a going concern. Borrowers
may be willing to pay shared auditors a premium to signal high quality financial statements to
obtain financing. Shared auditors may also have more insight in assessing the borrower’s ability
to secure financing, which improves the accuracy of the going concern opinion for firms that
subsequently file for bankruptcy.
Both specialist variables are significantly (p < 0.05) positively associated with audit fees.
Most other control variables are consistent with predictions. Size, losses, number of foreign
segments, going concern opinions, and December year-ends are all positively associated with
audit fees while current ratio and return on assets are negatively associated with audit fees.
Overall, these results are consistent with the prediction that borrower audit fees are positively
associated with shared auditors.
33

V. SUPPLEMENTAL ANALYSES

Cross-Sectional Analyses
Commercial Loan Portfolio
A commercial bank’s portfolio includes commercial and industrial loans, commercial real
estate loans, and consumer loans (i.e., mortgages, credit card loans, auto loans, student loans).
Portfolios vary between banks, and shared auditors are more likely to gather information relating
to the commercial portion of the bank’s portfolio. The potential for knowledge spillover is less
likely for banks with larger consumer loan or real estate loan portfolios.
The primary benefit of having a shared auditor is that information can be transferred
between engagements when auditing both the borrower and the lender in a commercial lending
relationship. To examine if information transfer is the mechanism for the benefits of having a
shared auditor, I perform cross-sectional tests that examines the relation between shared auditors
and audit quality partitioned on the lender’s commercial loan portfolio. Commercial loan
portfolio is based on the lender’s total commercial and agricultural loans divided by total loans.
Observations are split into low (high) commercial loan portfolio groups based on whether the
lender’s proportion of commercial loans is less (greater) than 20%, which is the approximate
median value of commercial loan portfolios in my sample and consistent with the average
proportion of commercial loans for the average bank in Bhat, Lee, and Ryan (2014). Tables 6
and 7 reports the results for lender audit quality and lender audit fees, respectively.
Table 6 reports that the association between shared auditors and loan loss provision
validity with a partition on commercial loans. The coefficient on the interaction between
LSHAREAUDOFF and LLP is positive and significantly in the subsample of high commercial
34
loans in the full sample (p<0.05) and matched sample (p<0.1). This result indicates that the
positive association between shared auditors and loan loss provision validity is greater for
lenders with larger commercial loan portfolios.
Table 7 reports the association between shared auditors and lender audit fees with a
partition on commercial loans. Results indicate that the positive association between shared
auditors and borrower audit fees is driven by firms with larger commercial loan portfolios. The
association between shared auditors and audit fees is not significant for lenders with smaller
commercial loan portfolios. A test for difference in coefficients between the low and high
commercial loan subsamples is significantly different (p<0.1) in both the full and propensity
score matched samples (untabulated). This result indicates that the positive association between
shared auditors and lender audit fees is primarily found in firms with larger commercial loan
portfolios. Overall, these results suggest that the effects of shared auditors accrue to the lender
only when the lender has a larger portfolio of commercial loans.
Lender Loan Importance
Shared auditors are more likely to develop knowledge about the borrower if the
borrower’s loan was reviewed as part of the lender’s audit. According to FASB ASC 310-10-35-
14, auditors should consider materiality and management reports of total loan amounts by
borrower when identifying loans for evaluation. Since the borrower’s loan is more likely to be
selected for evaluation if the borrower’s loan is significant to the lender, the development and
transfer of information by shared auditors is more likely for larger loans since they are more
important to the lender. To get a better understanding of the mechanism behind the effects of a
shared auditor, I reexamine the borrower analyses with a partition on loan importance. Loan
importance is split into high (low) importance based on whether the borrower’s total loans scaled
35
by the lender’s average loans are above (below) the median. Tables 8 and 9 report the results for
borrower audit quality and borrower audit fees, respectively.
Table 8 reports the going concern model for firms that subsequently filed for bankruptcy
partitioned on loan importance. The results indicate a significant (p < 0.01) positive association
for going concern reporting for borrowers with more important loans and a positive but non-
significant association for borrowers with less important loans in both the full and distressed
samples. It is interesting to note that the coefficients on leverage are insignificant for firms with
smaller loans but significant (p < 0.05) for firms with larger loans. Although TOTLEV is still
positively associated with a going concern, CLEV is negatively associated with a going concern.
This suggests that obtaining short-term financing, possibly through the lender, mitigates a going
concern opinion. Cash flow measures and obtaining financing are more important determinants
of a going concern opinion for these extremely distressed firms; standard book measures of
distress, such as bankruptcy or previous losses, are less likely to determine a going concern
opinion. These results indicate that the association between shared auditors and going concern
reporting is greater for borrowers with larger loans.
Table 9 reports the results of the effects of shared auditors on borrower audit fees
partitioned by loan importance. The significant positive association between shared auditors and
audit fees is exclusively in the subsample of firms with larger loans. The association between
shared auditors and borrower audit fees is non-significant or negative for smaller loans. A test
for difference in coefficients between the low and high loan importance subsamples is
significantly different (p<0.01) (untabulated). Based on the full (propensity-score matched)
sample, results indicate that borrowers pay a 13.2% (11.3%) or $214k ($183k) fee premium for
shared auditors when they have larger loans. These results are consistent with the prediction that
36
the association between shared auditors and borrower audit fees is greater for borrowers with
larger loans. Overall, these results suggest that the effects of shared auditors accrue to the
borrower only when the borrower’s loan is larger or more important to the lender.
Alternate Measure of Lender Shared Auditor
I use a binary variable to define a shared auditor in both the lender and borrower analyses
for consistency. However, the shared auditor effect may vary between lenders because each
lender has multiple borrowers. In supplemental analyses, I examine the lender shared auditor
effects using several alternate measures of lender shared auditor.
Continuous Measure
I use a continuous measure of lender shared auditor to examine the effect of each shared
auditor relationship. I replicate the primary lender analyses with LSHAREAUDOFFQ defined as
the number of lender shared office relationships.
Table 10, Panel A, Column (1) presents the results of the lender loan loss provision
validity analyses. The primary coefficient of interest is β
2
or the interaction between
LSHAREAUDOFF and LLP. There is a significantly positive association (p-value < 0.01)
between this variable and loan charge-offs. This indicates that shared auditors have a stronger
association between accrual and operating activity relative to different auditors or offices. These
results are consistent with the prediction that shared auditors are positively associated with the
validity of loan loss provisions.
Table 10, Panel B presents the results of the lender audit fee analyses. The results
indicate that LSHAREAUDOFFQ is positively and significantly (p < 0.05) associated with lender
audit fees and indicates that shared auditors increase audit fees by 2% for every shared auditor
37
relationship, holding all other factors constant. These results are consistent with the prediction
that shared auditors are positively associated with lender audit fees.
Materiality Measure
Shared auditor relationships are likely more meaningful to the lender if the balances
audited by shared auditors are material. I approximate materiality of shared auditor loans to the
lender audit using an indicator variable equal to one if the total value of shared auditor loans are
greater than or equal to five percent of the lender’s pre-tax income before loan loss provision,
and zero otherwise.
Table 10, Panel A, Column (2) presents the results of the lender loan loss provision
validity analyses. The primary coefficient of interest is β
2
or the interaction between
LSHAREAUDOFF and LLP. There is a significantly positive association (p-value < 0.05)
between this variable and loan charge-offs. The magnitude of the coefficient is in line with the
main results in Table 2, Panel C since the shared auditor is measured with an indicator variable
in both these models.
Table 10, Panel B presents the results of the lender audit fee analyses. The results
indicate that LSHAREAUDOFF is positively and significantly (p < 0.05) associated with lender
audit fees. The magnitude of the coefficient is in line with the primary results in Table 4, Panel
C. Overall, the results with a materiality measure of shared auditors is consistent with the
measure used in the main analysis.
Proportion Measure
I use an additional proportion measure to approximate the significance of shared auditor
loans to the lender. I replicate the primary lender analyses with LSHAREAUDOFF defined as the
total value of shared auditor loans divided by the lender’s total assets.
38
Table 10, Panel A, Column (3) presents the results of the lender loan loss provision
validity analyses. The primary coefficient of interest is the interaction between
LSHAREAUDOFF and LLP. There is a significantly positive association (p-value < 0.05)
between β
2
and loan charge-offs. The results are consistent with the primary findings and provide
support for the hypothesis that shared auditors are positively associated with the validity of loan
loss provisions.
Table 10, Panel B presents the results of the lender audit fee analyses. The results
indicate that LSHAREAUDOFF is positively and significantly (p < 0.05) associated with lender
audit. These results are consistent with the primary and alternate supplemental measures of
shared auditors and are consistent with the prediction that shared auditors are positively
associated with lender audit fees.
National Shared Auditor
I focus on the effects of shared auditor offices in this study because the mechanism for
information transfer is more likely to occur at the office level. Also, the effects of shared auditors
has previously been found to occur mainly at the office or city level (Dhaliwal et al., 2015;
Johnstone et al., 2014). However, shared auditor effects may occur at the national level (shared
auditor different office) through the national consultation group, knowledge management
databases, national training sessions, or the use of national audit teams.
To examine the effect of national shared auditors, I rerun all of the analyses for loan loss
provision validity, going concern reporting, and audit fees for the lender and borrower using
shared auditors from different offices as the primary variable of interest (untabulated). Results
for most analyses are not significant with the exception of the borrower audit fees. Shared
national auditors are significantly positively associated with borrower audit fees. Further, this
39
association exists only for loans that are more important to the lender. This suggests that
knowledge gained from shared auditors may be captured at the national level via identified risk
factors that the auditor factors into the pricing of the borrower’s audit. However, audit fees are a
noisy measure of audit quality and it is difficult to disentangle whether the premium reflects
increased effort or a risk premia. The results of these analyses suggest that the transfer of
knowledge content for shared auditors is primarily concentrated at the office level but not the
national level.
Geography Effects
It is possible that the information spillover occurs between the parties because of the
proximity of the auditor to the lender (borrower) (Audretsch & Feldman, 1996; Audretsch &
Stephan, 1996; Kedia & Rajgopal, 2011). I create a geography variable that is equal to one if the
headquarters of the lender (borrower) is in the same city as the auditor office. I incorporate this
geography variable into the tests and rerun Models (1), (2), (3) and (4). Table 11 presents the
results for the primary variable of interest. I find that the association between shared auditors and
the various measures of audit quality are positive and significant even with the control for
geography effects. This suggests that the knowledge spillover via shared auditors is incremental
to geography effects.
Falsification Tests
Knowledge spillover from auditing related firms in a banking relationship likely
provides the shared auditor information about downside risks. To provide additional evidence
that information transfers are the mechanism for the increase in audit quality, I examine the
relationship between shared auditors and alternate measures of audit quality that are not expected
to be affected by the knowledge spillover from auditing firms in a banking relationship.
40
Discretionary provisions or accruals are commonly used measures of audit quality that is less
likely to be affected by information spillover in this setting; hence, I do not expect that shared
auditors are associated with discretionary provisions or accruals. In additional analyses I
examine the association between shared auditors and discretionary loan loss provisions (Ahmed,
Takeda, & Thomas, 1999; Beatty & Liao, 2014; Liu & Ryan, 2006) for the lender and absolute
discretionary accruals (J. R. Francis & Michas, 2012) for the borrower.
For the lender, I estimate loan loss provisions based on Liu and Ryan (2006) and use the
residual from this model as the discretionary loan loss provision (untabulated). I do not find a
significant association between the shared auditor variable and discretionary loan loss provisions.
For the borrower, I estimate discretionary accruals based on the prior literature (J. R. Francis &
Michas, 2012). Consistent with the lender results, I do not find a significant association between
shared auditors and borrower discretionary accruals (untabulated).
Overall, I find that shared auditors are not significantly associated with either alternate
measure of audit quality. This provides additional support that the association between shared
auditors and audit quality are due to information spillover of client-specific knowledge; shared
auditors are not associated with general information transfers.

VI. CONCLUSION

This paper investigates whether an auditor develops client specific knowledge when
auditing both the borrower and lender and how this affects audit quality. I find that shared
auditors are positively associated with loan loss validity for the lenders, going concern reporting
accuracy for the borrower, and audit fee premiums for the lender and the borrower. I find that the
41
association between shared auditors and the various audit quality measures are only present for
lenders with larger commercial loan portfolios and borrowers with loans that are important to the
lender, which are circumstances where the banking relationships are relevant. These findings are
consistent with an improvement in audit quality for both the borrower and lender with a shared
auditor.
This study contributes to the literature by exploring a potential channel of auditor
competency. Prior literature on the supply of audit quality has primarily focused on auditor
characteristics such as size (i.e. BigN, small auditor) and industry specialization (DeFond &
Zhang, 2014). This study examines whether auditor knowledge and expertise developed while
auditing one party (i.e. lender or borrower) in a banking relationship can be transferred to the
other party. This is a unique form of auditor knowledge because these parties are in different
industries with different risks.
Another contribution of this study is that it provides archival support that auditor
knowledge may be transferrable across client firms and industries. By examining the effects of
shared auditors in banking relationships, the findings of this study shows that client specific
knowledge developed for one engagement are transferrable to other engagements in different
industries when there are common elements.
This study also extends the literature on shared auditors and suggests that shared auditors
can be beneficial for both parties involved. Prior research has examined the effects of shared
auditors on only one party or found that shared auditors benefit one party at the expense of the
other. This study shows that shared auditors increases audit quality for both the lender and the
borrower and the auditor earns a fee premium. This suggests shared auditors in banking
relationships benefit both the lender and the borrower.
42
The study can also inform regulators and practitioners about the mechanisms of auditor
competency, which audit firms can find useful in staff training and development. The findings
from this study suggest that auditors may develop key competencies by working on related
engagements outside of their industries. Working on related engagements can improve auditor
competencies in both fields and is incremental to the competencies gained from industry
experience.
This paper focuses on the effect of shared auditors on audit quality. Future research can
extend this study by examining whether common auditors affect the banking relationship
between the lender and borrower through lending decisions or the borrower’s cost of debt.
Barath et al. (2008) and Francis et al. (2005) examine how the borrower’s accounting quality
mitigates information asymmetry between the borrower and lender and find that it reduces the
interest rates charged to borrowers. This extension can examine whether the auditor plays a role
in reducing information asymmetry between the lender and borrower. This will provide
additional evidence on the effect of shared auditors and the potential spillover effect of auditor
expertise and audit quality on other stakeholders.

43
REFERENCES

Ahmed, A. S., Takeda, C., & Thomas, S. (1999). Bank loan loss provisions: a reexamination of
capital management, earnings management and signaling effects. Journal of Accounting
and Economics, 28(1), 1–25.
Altamuro, J., & Beatty, A. (2010). How does internal control regulation affect financial
reporting? Journal of Accounting and Economics, 49(1), 58–74.
Aobdia, D. (2015). Proprietary information spillovers and supplier choice: evidence from
auditors. Review of Accounting Studies, 1–36. http://doi.org/10.1007/s11142-015-9327-x
Arnold, V., Collier, P. A., Leech, S. A., & Sutton, S. G. (2001). The impact of political pressure
on novice decision makers: are auditors qualified to make going concern judgements?
Critical Perspectives on Accounting, 12(3), 323–338.
Audretsch, D. B., & Feldman, M. P. (1996). R&D Spillovers and the Geography of Innovation
and Production. The American Economic Review, 86(3), 630–640.
Audretsch, D. B., & Stephan, P. E. (1996). Company-Scientist Locational Links: The Case of
Biotechnology. The American Economic Review, 86(3), 641–652.
Ball, R. (2009). Market and Political/Regulatory Perspectives on the Recent Accounting
Scandals. Journal of Accounting Research, 47(2), 277.
Balsam, S., Krishnan, J., & Yang, J. S. (2003). Auditor industry specialization and earnings
quality. Auditing: A Journal of Practice & Theory, 22(2), 71–97.
Beatty, A., & Liao, S. (2013). Financial Accounting in the Banking Industry: A Review of the
Empirical Literature. Available at SSRN 2346752. Retrieved from
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2346752
Beatty, A., & Liao, S. (2014). Financial accounting in the banking industry: A review of the
empirical literature. Journal of Accounting and Economics, 58(2–3), 339–383.
http://doi.org/10.1016/j.jacceco.2014.08.009
Bell, T. B., Landsman, W. R., & Shackelford, D. A. (2001). Auditors’ Perceived Business Risk
and Audit Fees: Analysis and Evidence. Journal of Accounting Research, 39(1), 35–43.
http://doi.org/10.1111/1475-679X.00002
Bharath, S., Sunder, J., & Sunder, S. (2008). Accounting Quality and Debt Contracting. The
Accounting Review, 83(1), 1.
Bharath, S. T., Dahiya, S., Saunders, A., & Srinivasan, A. (2009). Lending Relationships and
Loan Contract Terms. Review of Financial Studies, hhp064.
http://doi.org/10.1093/rfs/hhp064
Bhat, G., Lee, J. A., & Ryan, S. G. (2014). Using Loan Loss Indicators by Loan Type to Sharpen
the Evaluation of the Determinants and Implications of Banks’ Loan Loss Accruals
(SSRN Scholarly Paper No. ID 2490670). Rochester, NY: Social Science Research
Network. Retrieved from http://papers.ssrn.com/abstract=2490670
Bonner, S. E. (2008). Judgment and decision making in accounting. Prentice Hall.
Carson, E., Fargher, N. L., Geiger, M. A., Lennox, C. S., Raghunandan, K., & Willekens, M.
(2012). Audit reporting for going-concern uncertainty: A research synthesis. Auditing: A
Journal of Practice & Theory, 32(sp1), 353–384.
Cashell, J. D., & Fuerman, R. D. (1995). The CPA’s Responsibility for Client Information. The
CPA Journal, 65(9), 54.
44
Chava, S., & Roberts, M. R. (2008). How Does Financing Impact Investment? The Role of Debt
Covenants. The Journal of Finance, 63(5), 2085–2121. http://doi.org/10.1111/j.1540-
6261.2008.01391.x
Craswell, A. T., Francis, J. R., & Taylor, S. L. (1995). Auditor brand name reputations and
industry specializations. Journal of Accounting and Economics, 20(3), 297–322.
http://doi.org/10.1016/0165-4101(95)00403-3
DeFond, M. L., Raghunandan, K., & Subramanyam, K. R. (2002). Do non–audit service fees
impair auditor independence? Evidence from going concern audit opinions. Journal of
Accounting Research, 40(4), 1247–1274.
DeFond, & Zhang. (2014). A review of archival auditing research. Journal of Accounting and
Economics, 58(2–3), 275–326. http://doi.org/10.1016/j.jacceco.2014.09.002
Dhaliwal, D. S., Lamoreaux, P. T., Litov, L. P., & Neyland, J. B. (2015). Shared auditors in
mergers and acquisitions. Journal of Accounting and Economics.
http://doi.org/10.1016/j.jacceco.2015.01.005
Dopuch, N., & Simunic, D. (1982). Competition in Auditing: An Assessment", Fourth
Symposium on Auditing Research, University of Illinois.
Federal Reserve Bank of New York. (2014). CRSP-FRB Link. Retrieved October 15, 2014, from
http://www.newyorkfed.org/research/banking_research/datasets.html
Ferguson, A., Francis, J. R., & Stokes, D. J. (2003). The effects of firm-wide and office-level
industry expertise on audit pricing. The Accounting Review, 78(2), 429–448.
Fields, L. P., Fraser, D. R., & Wilkins, M. S. (2004). An investigation of the pricing of audit
services for financial institutions. Journal of Accounting and Public Policy, 23(1), 53–77.
Francis, J., LaFond, R., Olsson, P., & Schipper, K. (2005). The market pricing of accruals
quality. Journal of Accounting and Economics, 39(2), 295–327.
Francis, J. R., & Michas, P. N. (2012). The contagion effect of low-quality audits. The
Accounting Review, 88(2), 521–552.
Francis, J. R., Reichelt, K., & Wang, D. (2005). The Pricing of National and City ‐Specific
Reputations for Industry Expertise in the U.S. Audit Market. The Accounting Review,
80(1), 113–136. http://doi.org/10.2308/accr.2005.80.1.113
GAO. (1991). Failed Banks: Accounting and Auditing Reforms Urgently Needed. Retrieved
October 10, 2014, from http://www.gao.gov/products/afmd-91-43
GAO. (1994). Depository Institutions: Divergent Loan Loss Methods Undermine Usefulness of
Financial Reports. Retrieved October 10, 2014, from
http://www.gpo.gov/fdsys/pkg/GAOREPORTS-AIMD-95-8/html/GAOREPORTS-
AIMD-95-8.htm
Geiger, M. A., Raghunandan, K., & Rama, D. V. (2005). Recent Changes in the Association
between Bankruptcies and Prior Audit Opinions. Auditing, 24(1), 21–35.
Johnstone, K. M., Li, C., & Luo, S. (2014). Client-Auditor Supply Chain Relationships, Audit
Quality, and Audit Pricing. AUDITING: A Journal of Practice & Theory, 33(4), 119–
166. http://doi.org/10.2308/ajpt-50783
Kanagaretnam, K., Krishnan, G., & Lobo, G. (2010). An Empirical Analysis of Auditor
Independence in the Banking Industry. The Accounting Review, 85(6), 2011.
Kedia, S., & Rajgopal, S. (2011). Do the SEC’s enforcement preferences affect corporate
misconduct? Journal of Accounting & Economics, 51(3), 259.
Krishnan, G. V. (2003). Does Big 6 auditor industry expertise constrain earnings management?
Accounting Horizons, 17, 1–16.
45
Landsman, W., Nelson, K., & Rountree, B. (2009). Auditor Switches in the Pre- and Post-Enron
Eras: Risk or Realignment? The Accounting Review, 84(2), 531.
Leuz, C., Triantis, A., & Wang, T. (2008). Why do firms go dark? Causes and economic
consequences of voluntary SEC deregistrations. Journal of Accounting & Economics,
45(2/3), 181.
Liu, C.-C., & Ryan, S. G. (2006). Income Smoothing over the Business Cycle: Changes in
Banks’ Coordinated Management of Provisions for Loan Losses and Loan Charge-Offs
from the Pre-1990 Bust to the 1990s Boom. The Accounting Review, 81(2), 421–441.
McAllister, B., & Cripe, B. (2008). Improper Release of Proprietary Information. The CPA
Journal, 78(3), 52–55.
Menon, K., & Williams, D. (2010). Investor Reaction to Going Concern Audit Reports. The
Accounting Review, 85(6), 2075.
Office of the Comptroller of the Currency. (1998, May). Comptroller’s Handbook: Allowance
for Loan and Lease Losses. Retrieved October 10, 2014, from
http://www.occ.gov/publications/publications-by-type/comptrollers-handbook/alll.pdf
Raghunandan, K., & Rama, D. v. (1995). Audit Reports for Companies in Financial Distress:
Before and After SAS No. 59. Auditing, 14(1), 50–63.
Reichelt, K., & Wang, D. (2010). National and Office-Specific Measures of Auditor Industry
Expertise and Effects on Audit Quality. Journal of Accounting Research, 48(3), 647.
Reynolds, J. K., & Francis, J. R. (2000). Does size matter? The influence of large clients on
office-level auditor reporting decisions. Journal of Accounting and Economics, 30(3),
375–400. http://doi.org/10.1016/S0165-4101(01)00010-6
Shleifer, A., & Vishny, R. W. (2010). Fire Sales in Finance and Macroeconomics (Working
Paper No. 16642). National Bureau of Economic Research. Retrieved from
http://www.nber.org/papers/w16642
Thibodeau, J. C. (2003). The Development and Transferability of Task Knowledge. AUDITING:
A Journal of Practice & Theory, 22(1), 47–67. http://doi.org/10.2308/aud.2003.22.1.47
Venuti, E. K. (2004). The Going-Concern Assumption Revisited: Assessing a Company’s Future
Viability. The CPA Journal, 74(5). Retrieved from
http://www.nysscpa.org/cpajournal/2004/504/essentials/p40.htm
Watts, R., & Zimmerman, J. (1982). Auditors and the Determination of Accounting Standards.
Retrieved from
https://urresearch.rochester.edu/institutionalPublicationPublicView.action?institutionalIte
mId=4512&versionNumber=1
Zmijewski, M. E. (1984). Methodological issues related to the estimation of financial distress
prediction models. Journal of Accounting Research, 59–82.

46
FIGURE 1
SHARED AUDITORS

Borrower A
Lender A Lender B
Auditor
Office 1
Borrower C
Borrower B
Lender C
Auditor
Office 2
Auditor
Office 3

This figure depicts a shared auditor office in a banking relationship (“shared auditor”). Lender A
and Borrower A have a shared auditor since they are both audited by Auditor Office 1. Borrower
B, Lender B, Borrower C, and Lender C have different auditors or the same auditors from
different offices. In lender analyses, Lender A is in the treatment group while Lender B and
Lender C are in the control group. In borrower analyses, Borrower A is in the treatment group
while Borrower B and Borrower C are in the control group.

47
APPENDIX A: VARIABLE DEFINITIONS

Variable Definition Source
LSHAREAUDOFF Lender with shared auditor office with an indicator variable equal to one if the
lender has greater than the median number of shared auditor office relationships,
and zero otherwise. Shared auditor office is when the lender and borrower received
audit opinions from the same auditor and city.
Audit Analytics
SHAREAUDOFF Borrower with shared auditor office with an indicator variable equal to one if the
lender and borrower received audit opinions from the same auditor and city, and
zero otherwise.
Audit Analytics
LSPECIALIST Lender auditor specialist based on the market share approach with an indicator
variable equal to one if the auditor has the largest market share of audit fees in a
two-digit SIC category year, and zero otherwise.
Audit Analytics
SPECIALIST Borrower auditor specialist based on the market share approach with an indicator
variable equal to one if the auditor has the largest market share of audit fees in a
two-digit SIC category year, and zero otherwise.
Audit Analytics

Lender Variables
CAPRATIO Total risk-adjusted capital ratio. FRB Chicago
CHGOFF Net charge-offs during year t+1 deflated by beginning total assets. FRB Chicago
CHGOFFCY Net charge-offs during year t deflated by beginning total assets. FRB Chicago
COMMLOAN Total commercial and agricultural loans divided by total loans. FRB Chicago
CONSUMERLOAN Total consumer loans divided by total loans. FRB Chicago
EFFICIENCY Efficiency ratio measured by total operating expenses deflated by total revenue. FRB Chicago
EXEMPT Exempted banks with an indicator variable that equals 1 if a bank has less than $500
million in total assets during the years 2000 through 2003 or less than $500 million
in total assets and public float less than $75 million for the year 2004 or less than $1
billion in total assets and public float less than $75 million for the years 2005 and
2006, and 0 otherwise.
FRB Chicago,
Audit Analytics
INTANG Intangible assets divided by total assets. FRB Chicago
LAUDITFEES Lender audit fees defined as the natural logarithm of audit fees. Audit Analytics
LLOSS Loss defined as an indicator variable that equals 1 if net income is negative, and
zero otherwise.
FRB Chicago
LLP Loan loss provision during year t, scaled by beginning total assets. FRB Chicago
LLSIZE Size defined as the log of total assets in millions at the beginning of year t. FRB Chicago
LSIZE Natural log of total assets. FRB Chicago
NONACC Non-performing loans at the end of year t, scaled by beginning total assets. FRB Chicago
NONPERFORM Ratio of non-performing loans divided by lagged total loans. FRB Chicago
RELOAN Total real estate loans divided by total loans. FRB Chicago
SECURITIES Defined as one less total securities, scaled by total assets. FRB Chicago
SENSITIVE Rate-sensitive assets minus rate-sensitive liabilities. FRB Chicago
TOTLOANS Total loans, scaled by total assets. FRB Chicago

48
APPENDIX A: VARIABLE DEFINITIONS (continued)

Variable Definition Source
Borrower Variables
AGE Natural logarithm of firm age, where firm age equals the number of years a firm’s
financial data is available from Compustat.
Compustat
AUDITORSWITCH Indicator variable that equals one when the firm switches auditors in the current
year, and zero otherwise.
Audit Analytics
BAUDITFEES Borrower audit fees defined as the natural logarithm of audit fees. Audit Analytics
BLLOSS Prior year loss with an indicator variable equals to one when net income is negative,
and zero otherwise.
Compustat
BLOSS Current year loss with an indicator variable equals to one when a firm’s net income
is negative, and zero otherwise.
Compustat
CFO Operating cash flow scaled by total assets. Compustat
CLEV Change in leverage from year t-1 to year t.
CURR Current ratio defined as current assets scaled by current liabilities. Compustat
DECYE Indicator variable that equals one if the firm has a December fiscal year-end, and
zero otherwise.

FUTUREFINANCE Indicator variable equal to one when the firm issues equity or long-term debt in the
subsequent year, and zero otherwise.
Compustat
INVESTMENTS Short- and long-term investment securities (including cash and cash equivalents)
scaled by total assets at year-end.
Compustat
LEV Leverage defined as the sum of current and long-term debt scaled by average total
assets.
Compustat
LOANIMP Importance of the loan to the lender defined as an indicator variable equal to one if
the loan size is greater than the median loan size, and zero otherwise. Loan size is
defined as the total loan scaled by the average loan outstanding that year.
DealScan, FRB
Chicago
MB Market value of equity scaled by book value of equity at the end of year t. Compustat
NUMFOREIGNSEGMENTS # of geographic segments defined as the natural logarithm of one plus the number of
geographic segments.
Compustat
NUMOPESEGMENTS # of operating segments defined as the natural logarithm of one plus the number of
operating segments.
Compustat
OPINION Borrower going concern optinion with an indicator variable equal to one if the
firm's auditor issued a going concern opinion, and zero otherwise.
Audit Analytics
QUICK Quick ratio defined as current assets (excluding inventories) scaled by current
liabilities.
Compustat
RISKY Indicator variable equal to one if the company operated in a risky industry (SIC
codes 2833, 2836, 3570, 3577, 3600, 3674, 7372, 7379, 8731, and 8734).
Compustat
RLAG Number of days between fiscal year-end and the opinion date. Audit Analytics
ROA Return on assets defined as net income scaled by average total assets. Compustat
SHARE_ISSUE Indicator variable equal to one when a firm issued equity in the fiscal year, and zero
otherwise.
Compustat
SIZE Size defined as the natural logarithm of total assets. Compustat
TOTLEV Total liabilities over total assets at the end of the fiscal year (levgc). Compustat
ZMIJEVSKI Probability of bankruptcy from Zmijevski (1984). Zmijevski=[-4.336-4.513*(Net
Income/Total Assets)+5.679*(Total Liabilities/Total Assets)+0.004*(Current
Assets/Current Liabilities)].
Compustat
49
TABLE 1
SAMPLE SELECTION

Panel A: Sample Selection
Bank
relationships
Borrower
firm-years
Lender
firm-years
DealScan Database: Lead banking relationships between 2000-2012 21,824 88,310 1,575
Less: observations not matched with Compustat and CRSP data (6,787) (9,770) (100)
Banking relationships with WRDS data 15,037 78,540 1,475
Less: ADRs and borrowers incorporated outside the U.S. or Canada (140) (731) -
Less: financial (SIC 6000-6999) or utility (SIC 4900-4999) firms (1,996) (11,146) (38)
Less: lenders without Audit Analytics data (2,390) (10,894) (442)
Banking relationships with Audit Analytics data 10,511 55,769 995
Less: lenders with non-BigN auditors (937) (4,890) (126)
Banking relationships with BigN auditors 9,574 50,879 869
Less: observations without FR-Y9C data (1,882) (7,686) (400)
Observations with borrower and lender data 7,692 43,193 469

Final samples after removing observations with missing control variables:
Lender audit quality sample (Table 2) 406
Borrower audit quality sample (Table 3) 26,130
Lender audit fees sample (Table 4) 389
Borrower audit fees sample (Table 5) 39,415

Panel B: Shared Auditor Offices by Year

Borrower
firm-years

Lender
firm-years

Shared auditor office
firm-years

Observations %

Observations %

Observations %
2000 3,214 7.36

47 10.02

123 10.39
2001 3,307 7.63

56 11.94

110 9.29
2002 3,430 7.86

52 11.09

130 10.98
2003 3,164 7.25

50 10.66

119 10.05
2004 3,489 8.1

42 8.96

85 7.18
2005 3,385 7.86

31 6.61

81 6.84
2006 3,298 7.67

29 6.18

78 6.59
2007 3,157 7.3

31 6.61

87 7.35
2008 3,063 7.11

25 5.33

73 6.17
2009 3,236 7.53

29 6.18

75 6.33
2010 3,380 7.87

28 5.97

73 6.17
2011 3,649 8.5

25 5.33

76 6.42
2012 3,421 7.96

24 5.12

74 6.25
Total 43,193 469 1,184

Panel A provides details on the sample selection process. Panel B reports details of the subsamples by fiscal year.

50

TABLE 2
SHARED AUDITORS AND LENDER LOAN LOSS PROVISION VALIDITY

Panel A: Univariate statistics partitioned on shared auditor offices

Different Auditors or Offices
(N=226)
Shared Auditor Offices
(N=180) Difference
Variables Mean Median Std Dev Mean Median Std Dev Mean Median
Lender Quality Variables
CHGOFF
0.008 0.004 0.009 0.010 0.008 0.009 ** ***
LLP 0.006 0.003 0.008 0.006 0.004 0.006

***
LLSIZE
17.148 17.253 1.368 18.818 18.763 1.694 *** ***
LSPECIALIST
0.248 0.000 0.433 0.317 0.000 0.466
NONACC
0.009 0.006 0.009 0.012 0.008 0.011 *** ***
LSHAREAUDOFFQ
0.000 0.000 0.000 6.128 3.000 7.261 *** ***

Panel B: Test of Covariate Balance

Shared
Auditor Firms
Control
Firms Difference in Mean

Variable Mean Mean t-stat p-value

Before PSM

LLP
0.007 0.006 1.80 0.073
*
LLSIZE
19.667 17.394 12.98 0.000
***
LSPECIALIST
0.393 0.246 2.77 0.006
***
NONACC
0.014 0.010 3.21 0.001
***

After PSM

LLP
0.007 0.007 -0.33 0.745

LLSIZE
19.440 19.243 1.18 0.241

LSPECIALIST
0.316 0.215 1.44 0.152

NONACC
0.012 0.012 0.16 0.872

These tables report summary statistics for the sample of firms in banking relationships per the Dealscan database from 2000
through 2012. Panel A reports descriptive statistics for the subsample of firms with and without shared auditor offices.
Differences are and are based on t-statistics (z-statistics) for differences in the means (median) values between the different
auditor or offices and shared auditor offices samples. Panel B reports the covariate balance before and after propensity-score
matching. See the Appendix A for all variable definitions. All continuous variables are winsorized at the 1% and 99% levels. *,
**, and *** denote significance at the ten, five, and one percent levels, respectively, and are based on t-statistics (z-statistics)
for differences in the means (median) values between samples.

51
TABLE 2
SHARED AUDITORS AND LENDER LOAN LOSS PROVISION VALIDITY

CHGOFF
t+1
= α + β
1
LSHAREAUDOFF
t
+ β
2
LSHAREAUDOFF
t
*LLP
t
+ Controls
t
+ ϵ (1)

Panel C: Multivariate Analysis (1) (2) (3)
VARIABLES
Predicted
Sign Base Model Full Sample
PSM
Sample

LSHAREAUDOFF -0.000 -0.001
(-0.453) (-0.830)
LSHAREAUDOFF*LLP + 0.274*** 0.305**
(2.654) (2.757)
LSPECIALIST -0.002* -0.001
(-1.977) (-0.858)
LSPECIALIST*LLP + 0.241* 0.217
(1.779) (1.452)
LLP + 0.445*** 0.344*** 0.338***
(4.428) (3.716) (3.188)
LLSIZE + 0.002 0.001 0.001
(1.648) (0.863) (0.796)
NONACC + 0.117 0.157 -0.023
(1.212) (1.645) (-0.194)
CONSTANT -0.025 -0.012 -0.015
(-1.383) (-0.592) (-0.507)

Observations 406 406 173
Adjusted R-squared 0.753 0.771 0.804

This table presents the relationship between shared auditors and the validity of loan loss provisions, which is how
well loan loss provisions map into actual net charge-offs. Columns (1) presents the based model before the variables
of interest are added. Column (2) presents the analysis using the full sample. Column (3) presents the analysis using
a propensity-score matched sample. CHGOFF is the next period’s net loan charge-offs scaled by total assets. LLP is
loan-loss provision during year t scaled by beginning total assets. LSHAREAUDOFF equals one if the lender has
greater than the median number of shared auditor office relationships, and zero otherwise. LSPECIALIST is an
indicator variable that equals one if the lender’s auditor has the largest market share of audit fees in a two-digit SIC
category year, and zero otherwise. LLSIZE is the log of total assets at the beginning of year t. NONACC is non-
performing loans at the end of year t scaled by beginning total assets. All continuous variables are winsorized at the
1% and 99% levels. Year indicators control for year fixed effects. Standard errors are clustered by company. T-
statistics are reported in parentheses. *, **, and *** denote significance at the ten, five, and one percent levels,
respectively.

52
TABLE 3
SHARED AUDITORS AND BORROWER GOING CONCERN REPORTING

Panel A: Univariate statistics partitioned on shared auditor offices

Different Auditors or Offices (N=25,366) Shared Auditor Offices (N=764) Difference
Variables Mean Median Std Dev Mean Median Std Dev Mean Median
Borrower Quality Variables
AGE
2.970 2.944 0.732 3.258 3.526 0.749 *** ***
BANKRUPT
0.007 0.000 0.086 0.013 0.000 0.114 *** ***
BLLOSS
0.211 0.000 0.408 0.188 0.000 0.391 *** ***
CFO
0.099 0.098 0.078 0.108 0.110 0.070 *** ***
CLEV
0.001 -0.004 0.082 -0.001 -0.003 0.066 *** ***
FUTUREFINANCE
0.952 1.000 0.213 0.949 1.000 0.220
INVESTMENTS
0.126 0.082 0.133 0.138 0.108 0.122
LSPECIALIST
0.514 1.000 0.500 0.399 0.000 0.490 *** ***
MB
2.665 2.053 3.698 3.309 2.299 4.316 *** ***
OPINION
0.010 0.000 0.098 0.007 0.000 0.081 *** ***
RISKY
0.134 0.000 0.340 0.098 0.000 0.298
RLAG
7.527 7.616 1.116 7.279 7.483 1.160 *** ***
ROA
0.031 0.047 0.110 0.050 0.055 0.097 *** ***
SHARE_ISSUE
0.858 1.000 0.349 0.834 1.000 0.373 ** **
SHAREAUDOFF
0.000 0.000 0.000 1.000 1.000 0.000 *** ***
SIZE
7.438 7.389 1.700 7.804 7.564 1.945 *** ***
SPECIALIST
0.337 0.000 0.473 0.436 0.000 0.496 *** ***
TOTLEV
0.586 0.570 0.231 0.569 0.563 0.210
ZMIJEVSKI
0.190 0.070 0.262 0.161 0.054 0.233 *** ***

These tables report summary statistics for the sample of firms in banking relationships per the Dealscan database from 2000
through 2012. Panel A reports descriptive statistics for the subsample of firms with and without shared auditor offices.
Differences are and are based on t-statistics (z-statistics) for differences in the means (median) values between the different
auditor or offices and shared auditor offices samples. See the Appendix A for all variable definitions. All continuous variables
are winsorized at the 1% and 99% levels. *, **, and *** denote significance at the ten, five, and one percent levels,
respectively, and are based on t-statistics (z-statistics) for differences in the means (median) values between the different
auditor or offices and shared auditor offices samples.

53
TABLE 3 (continued)
SHARED AUDITORS AND BORROWER GOING CONCERN REPORTING

OPINION = α + β
1
SHAREAUDOFF + Controls + ϵ (2)
Panel B: (1) (2) (3) (4)
Multivariate Analysis Full Sample Distressed Sample
Predicted Subsequently Subsequently Subsequently Subsequently
VARIABLES Sign Bankrupt Non-Failing Bankrupt Non-Failing
SHAREAUDOFF +/- 2.367** -0.973 2.084** -0.951
(2.416) (-0.903) (2.133) (-0.900)
SPECIALIST +/- -0.216 -0.567*** -0.184 -0.423**
(-0.540) (-2.695) (-0.402) (-2.018)
LSPECIALIST +/- 0.313 0.060 0.266 0.016
(0.849) (0.336) (0.683) (0.089)
ZMIJEVSKI + 1.873 2.316*** 0.800 2.036***
(1.069) (4.413) (0.476) (3.997)
BLLOSS + -0.826 0.755*** -0.890 0.352
(-1.419) (3.393) (-1.385) (1.573)
AGE - -0.588* -0.055 -0.702** -0.166
(-1.926) (-0.390) (-2.079) (-1.165)
TOTLEV +/- -2.122 -0.043 -1.070 0.419
(-1.222) (-0.074) (-0.739) (0.787)
CLEV +/- -4.184 -0.321 -2.895 -0.586
(-1.160) (-0.362) (-0.766) (-0.734)
CFO - -3.161 -3.293*** -2.736 -1.274*
(-1.112) (-3.012) (-0.900) (-1.722)
RISKY +/- -2.805*** -0.434 -2.058** -0.549**
(-2.871) (-1.637) (-2.158) (-2.086)
SIZE - 0.282 -0.410*** 0.129 -0.476***
(1.621) (-5.953) (0.704) (-6.731)
ROA - -5.921** -3.250*** -3.234 -2.303***
(-2.310) (-4.806) (-1.081) (-3.381)
MB - 0.023 -0.029 0.035 -0.010
(0.249) (-1.568) (0.357) (-0.694)
SHAREISSUE - -1.085** -0.334 -1.685*** -0.054
(-2.468) (-1.563) (-3.294) (-0.239)
FUTUREFINANCE - -0.035 -0.624** -0.272 -0.343
(-0.067) (-2.419) (-0.502) (-1.251)
INVESTMENTS +/- 10.213*** -2.925*** 14.491*** -2.623***
(3.454) (-4.079) (3.392) (-3.831)
RLAG + 0.844*** 0.634*** 0.562*** 0.335***
(4.365) (8.660) (4.379) (6.874)
CONSTANT -7.576*** -7.380*** -3.483 -4.126***
(-3.086) (-7.758) (-1.632) (-5.194)

Observations 199 25,931 185 5,920
Pseudo R-squared 0.275 0.409 0.283 0.271
ROC Area 0.832 0.965 0.838 0.894

This table presents the relationship between shared auditors and going concern reporting. Subsequently bankrupt firms are firms that have filed for bankruptcy. Subsequently
non-failing firms are firms that have not filed for bankruptcy as of the end of 2013. Columns (1) and (2) presents the analysis using the full sample. Columns (3) and (4)
presents the analysis using a matched sample of distressed firms, which are firms with losses in the current year and negative cash flows from operations. OPINION is an
indicator variable equal to one the auditor issued a going concern opinion, and zero otherwise. SHAREAUDOFF is an indicator variable equal to one if the lender and
borrower received audit opinions from the same auditor office, and zero otherwise. SPECIALIST is an indicator variable that equals one if the borrower’s auditor has the
largest market share of audit fees in a two-digit SIC category year, and zero otherwise. LSPECIALIST is an indicator variable that equals one if the lender’s auditor has the
largest market share of audit fees in a two-digit SIC category year, and zero otherwise. ZMIJEWSKI is a bankruptcy measure with higher values indicating higher probability
of bankruptcy. BLLOSS is an indicator variable that equals one if the firm had a loss in the prior year, and zero otherwise. AGE is the log of number of years the company has
been publicly traded. TOTLEV is total liabilities over total assets at the end of the fiscal year. CLEV is the change in leverage during the year. CFO is operating cash flow
scaled by total assets. RISKY is an indicator variable equal to one if the company operated in a risky industry (SIC codes 2833, 2836, 3570, 3577, 3600, 3674, 7372, 7379,
8731, and 8734). SIZE is the log of total assets. ROA is net income scaled by average total assets. MB is the market value of equity scaled by book value of equity at the end of
year t. SHAREISSUE is an indicator variable equal to one when a firm issued equity in the fiscal year, and zero otherwise. FUTUREFINANCE is an indicator variable equal to
one when the firm issues equity or long-term debt in the subsequent year, and zero otherwise. INVESTMENTS are short- and long-term investment securities (including cash
and cash equivalents) scaled by total assets at year-end. RLAG is the number of days between the fiscal year-end and the opinion date. All continuous variables are winsorized
at the 1% and 99% levels. T-statistics are reported in parentheses. *, **, and *** denote significance at the ten, five, and one percent levels, respectively.
54
TABLE 4
SHARED AUDITORS AND LENDER AUDIT FEES

Panel A: Univariate statistics partitioned on shared auditor offices

Different Auditors or Offices
(N=221)
Shared Auditor Offices
(N=168) Difference
Variables Mean Median Std Dev Mean Median Std Dev Mean Median
Lender Fee
Variables
CAPRATIO
13.484 13.270 2.258 13.379 12.540 2.267
CHGOFFCY
0.374 0.293 0.295 0.452 0.442 0.256 ** ***
COMMLOAN
0.254 0.221 0.153 0.257 0.226 0.115
CONSUMERLOAN
0.091 0.060 0.101 0.133 0.129 0.079 *** ***
EXEMPT
0.348 0.000 0.478 0.321 0.000 0.468
INTANG
0.027 0.019 0.022 0.028 0.027 0.018

*
LAUDITFEES
14.442 14.445 1.218 15.752 15.636 1.471 *** ***
LLOSS
0.095 0.000 0.294 0.060 0.000 0.237
LSIZE
17.277 17.393 1.403 18.956 18.974 1.698 *** ***
LSPECIALIST
0.253 0.000 0.436 0.315 0.000 0.466
NONPERFORM
0.015 0.010 0.014 0.022 0.015 0.019 *** ***
RELOAN
0.552 0.599 0.192 0.480 0.497 0.147 *** ***
SECURITIES
0.350 0.695 0.990 0.654 0.721 0.408 *** **
SENSITIVE
0.230 0.221 0.141 0.267 0.254 0.141 *** **

55
TABLE 4 (continued)
SHARED AUDITORS AND LENDER AUDIT FEES

Panel B: Test of Covariate Balance

Shared Auditor
Firms
Control
Firms Difference in Mean
Variable Mean Mean t-stat p-value
Before PSM
CAPRATIO
13.379 13.484 -0.45 0.652
CHGOFFCY
0.010 0.008 2.72 0.007 ***
COMMLOAN
0.257 0.254 0.22 0.829
CONSUMERLOAN
0.133 0.091 4.44 0.000 ***
EXEMPT
0.321 0.348 -0.56 0.578
INTANG
0.028 0.027 0.69 0.492
LLOSS
0.060 0.095 -1.28 0.201
LSIZE
18.956 17.277 10.67 0.000 ***
LSPECIALIST
0.315 0.253 1.35 0.178
NONPERFORM
0.022 0.015 4.29 0.000 ***
RELOAN
0.480 0.552 -4.04 0.000 ***
SECURITIES
0.654 0.350 3.75 0.000 ***
SENSITIVE
0.267 0.230 2.62 0.009 ***

After PSM
CAPRATIO
13.288 13.487 -0.63 0.531
CHGOFFCY
0.008 0.008 0.41 0.679
COMMLOAN
0.278 0.279 -0.05 0.962
CONSUMERLOAN
0.104 0.118 -1.18 0.238
EXEMPT
0.398 0.442 -0.65 0.518
INTANG
0.028 0.025 1.13 0.259
LLOSS
0.074 0.068 0.17 0.864
LSIZE
18.045 17.897 0.92 0.357
LSPECIALIST
0.213 0.200 0.23 0.822
NONPERFORM
0.017 0.016 0.23 0.821
RELOAN
0.515 0.475 1.88 0.061 *
SECURITIES
0.633 0.614 0.27 0.787
SENSITIVE
0.277 0.282 -0.23 0.817

These tables report summary statistics for the sample of firms in banking relationships per
the Dealscan database from 2000 through 2012. Panel A reports descriptive statistics for the
subsample of firms with and without shared auditor offices. Differences are and are based on
t-statistics (z-statistics) for differences in the means (median) values between the samples.
Panel B reports the covariate balance before and after propensity-score matching. See the
Appendix A for all variable definitions. All continuous variables are winsorized at the 1%
and 99% levels. *, **, and *** denote significance at the ten, five, and one percent levels,
respectively.

56
TABLE 4 (continued)
SHARED AUDITORS AND LENDER AUDIT FEES
LAUDITFEE = α + β
1
LSHAREAUDOFF + Controls + ϵ (3)
Panel C: Multivariate Analysis (1) (2)
VARIABLES Predicted Sign Full Sample PSM Sample
LSHAREAUDOFF +/- 0.228** 0.265**
(2.006) (2.263)
LSPECIALIST + 0.316*** 0.315*
(3.238) (1.981)
LSIZE + 0.668*** 0.555***
(11.654) (6.881)
LLOSS + 0.137 0.252
(0.976) (1.445)
SECURITIES + -0.264** -0.185
(-2.149) (-0.920)
NONPERFORM + 4.705 12.312**
(1.196) (2.024)
CHGOFFCY + -1.320 -8.336
(-0.197) (-1.018)
COMMLOAN +/- 0.418 -0.036
(0.495) (-0.052)
CONSUMERLOAN +/- -0.228 -0.350
(-0.279) (-0.370)
RELOAN +/- -0.431 -0.732
(-0.650) (-1.130)
INTANG + 0.333 3.110
(0.126) (0.696)
CAPRATIO + 0.060** 0.117***
(2.573) (3.720)
EXEMPT - -0.359*** -0.012
(-3.236) (-0.084)
SENSITIVE - -0.309 -0.492
(-0.948) (-1.088)
CONSTANT 2.071 3.053*
(1.412) (1.683)

Observations 389 255
Adjusted R-squared 0.894 0.806
This table presents the relationship between shared auditors and lender audit fees. Column (1) presents the full sample. Column (2)
presents the analysis using a propensity-score matched sample. LAUDITFEE is the natural log of lender audit fees. LSHAREAUDOFF
equals one if the lender has greater than the median number of shared auditor office relationships, and zero otherwise. LSPECIALIST
equals one if the lender’s auditor has the largest market share of audit fees in a two-digit SIC category year, and zero otherwise. LSIZE
is the natural log of total assets. LLOSS equals one if net income is negative, and zero otherwise. SECURITIES is one less total
securities deflated by total assets. NONPERFORM is non-performing loans divided by lagged total loans. CHGOFFCY is net charge-
offs deflated by lagged total assets. COMMLOAN is total commercial and agricultural loans divided by total loans.
CONSUMERLOAN is total consumer loans divided by total loans. RELOAN is total real estate loans divided by total loans. INTANG is
intangible assets divided by total assets. CAPRATIO is total risk-adjusted capital ratio. EXEMPT equals one if a bank is exempt from
FDICIA and Section 404 of SOX, and zero otherwise. SENSITIVE is rate-sensitive assets minus rate-sensitive liabilities. All
continuous variables are winsorized at the 1% and 99% levels. Year indicators control for year fixed effects. Standard errors are
clustered by company. T-statistics are reported in parentheses. *, **, and *** denote significance at the ten, five, and one percent
levels, respectively.
57
TABLE 5
SHARED AUDITORS AND BORROWER AUDIT FEES

Panel A: Univariate statistics partitioned on shared auditor offices

Different Auditors or Offices
(N=38,329)
Shared Auditor Offices
(N=1,086) Difference
Variables Mean Median Std Dev Mean Median Std Dev Mean Median
Borrower Fee Variables
AUDITORSWITCH
0.042 0.000 0.200 0.031 0.000 0.174 *** **
BAUDITFEES
14.299 14.311 1.202 14.468 14.459 1.438 *** ***
BLOSS
0.242 0.000 0.428 0.198 0.000 0.399 *** ***
CURR
1.857 1.608 1.050 1.854 1.648 1.066
DECYE
0.689 1.000 0.463 0.703 1.000 0.457
LEV
0.324 0.280 0.242 0.283 0.233 0.227 *** ***
LSPECIALIST
0.515 1.000 0.500 0.431 0.000 0.495 *** ***
NUMFOREIGNSEGMENTS
0.728 0.693 0.708 0.838 0.693 0.652 *** ***
NUMOPESEGMENTS
1.039 1.099 0.611 1.080 1.386 0.654 * ***
OPINION
0.019 0.000 0.137 0.011 0.000 0.105 ** *
QUICK
1.343 1.138 0.873 1.282 1.111 0.836 ** *
ROA
0.025 0.042 0.113 0.046 0.055 0.111 *** ***
SHAREAUDOFF
0.000 0.000 0.000 1.000 1.000 0.000 *** ***
SIZE
7.539 7.514 1.683 7.785 7.491 1.887 *** ***
SPECIALIST
0.343 0.000 0.475 0.426 0.000 0.495 *** ***

58
TABLE 5 (continued)
SHARED AUDITORS AND BORROWER AUDIT FEES

Panel B: Test of Covariate Balance

Shared
Auditor Firms
Control
Firms
Difference in
Mean

Variable Mean Mean t-stat p-value

Before PSM

AUDITORSWITCH
0.030 0.042 -2.14 0.032
**
BLOSS
0.206 0.249 -3.32 0.001
***
CURR
1.850 1.847 0.09 0.930

DECYE
0.699 0.686 0.91 0.361

LEV
0.300 0.330 -4.09 0.000
***
LSIZE
19.877 20.541 -19.17 0.000
***
LSPECIALIST
0.436 0.518 -5.46 0.000
***
NUMFOREIGNSEGMENTS
0.825 0.721 4.93 0.000
***
NUMOPESEGMENTS
1.069 1.039 1.6 0.109

OPINION
0.017 0.024 -1.59 0.111

QUICK
1.279 1.335 -2.15 0.031
**
ROA
0.043 0.023 5.93 0.000
***
SIZE
7.710 7.506 4 0.000
***
SPECIALIST
0.416 0.339 5.37 0.000
***

After PSM

AUDITORSWITCH
0.030 0.032 -0.29 0.772

BLOSS
0.207 0.208 -0.05 0.960

CURR
1.852 1.839 0.29 0.773

DECYE
0.698 0.692 0.31 0.760

LEV
0.299 0.295 0.43 0.669

LSIZE
19.882 19.849 0.54 0.590

LSPECIALIST
0.437 0.443 -0.27 0.790

NUMFOREIGNSEGMENTS
0.825 0.791 1.19 0.234

NUMOPESEGMENTS
1.069 1.056 0.49 0.622

OPINION
0.017 0.018 -0.19 0.847

QUICK
1.281 1.267 0.41 0.683

ROA
0.043 0.040 0.6 0.546

SIZE
7.706 7.668 0.49 0.621

SPECIALIST
0.414 0.419 -0.23 0.818

These tables report summary statistics for the sample of firms in banking
relationships per the Dealscan database from 2000 through 2012. Panel A reports
descriptive statistics for the subsample of firms with and without shared auditor
offices. Differences are and are based on t-statistics (z-statistics) for differences in
the means (median) values between the samples. Panel B reports the covariate
balance before and after propensity-score matching. See the Appendix A for all
variable definitions. All continuous variables are winsorized at the 1% and 99%
levels. *, **, and *** denote significance at the ten, five, and one percent levels,
respectively.

59
TABLE 5 (continued)
SHARED AUDITORS AND BORROWER AUDIT FEES
BAUDITFEE = α + β
1
SHAREAUDOFF + Controls + ϵ (4)
Panel C: Multivariate Analysis (1) (2)
Variables Predicted Sign Full Sample PSM Sample
SHAREAUDOFF +/- 0.073* 0.065*
(1.654) (1.735)
SPECIALIST + 0.062*** 0.075**
(3.299) (2.331)
LSPECIALIST + 0.037*** 0.065**
(3.835) (2.299)
SIZE + 0.532*** 0.556***
(66.216) (43.839)
CURR - -0.075*** -0.091***
(-3.091) (-2.944)
QUICK - -0.004 0.001
(-0.140) (0.025)
ROA - -0.479*** -0.577***
(-6.456) (-3.131)
LEV + -0.150*** -0.307***
(-2.641) (-2.982)
BLOSS + 0.091*** 0.037
(4.427) (0.740)
NUMOPESEGMENTS + 0.029 0.001
(1.501) (0.028)
NUMFOREIGNSEGMENTS + 0.213*** 0.198***
(10.862) (5.233)
AUDITORSWITCH +/- -0.005 -0.058
(-0.169) (-0.799)
OPINION + 0.185*** 0.427***
(2.805) (3.643)
DECYE + 0.058** 0.076*
(2.194) (1.786)
CONSTANT 9.480*** 9.461***
(79.993) (67.378)

Observations 39,415 4,512
Adjusted R-squared 0.821 0.884

This table presents the relationship between shared auditors and borrower audit fees. Column (1) presents the analysis using the full sample. Column (2)
presents the analysis using a propensity-score matched sample. BAUDITFEE is the natural log of borrower audit fees. SHAREAUDOFF is an indicator
variable equal to one if the lender and borrower received audit opinions from the same auditor office, and zero otherwise. SPECIALIST is an indicator
variable that equals one if the borrower’s auditor has the largest market share of audit fees in a two-digit SIC category year, and zero otherwise and
LSPECIALIST is an indicator variable that equals one if the lender’s auditor has the largest market share of audit fees in a two-digit SIC category year,
and zero otherwise. SIZE is the natural log of assets. CURR is the ratio of current assets to current liabilities. QUICK is the ratio of current assets
excluding inventories to current liabilities. ROA is net income scaled by average total assets. LEV is the sum of current and long-term debt scaled by
average total assets. BLOSS is an indicator variable equal to one if the firm has negative net income in the current year, and zero otherwise.
NUMOPESEGMENTS is the natural logarithm of one plus the number of operating segments. NUMFOREIGNSEGMENTS is the natural logarithm of
one plus the number of foreign segments. AUDITORSWITCH is an indicator variable equal to one if the firm switched auditors for the current year, and
zero otherwise. OPINION is an indicator variable equal to one if the firm received a going concern opinion, and zero otherwise. DECYE is an indicator
variable equal to one if the firm has a December fiscal year-end, and zero otherwise. All continuous variables are winsorized at the 1% and 99% levels.
Year and industry indicators based on two-digit SIC codes control for year and industry fixed effects. Standard errors are clustered by company. T-
statistics are reported in parentheses. *, **, and *** denote significance at the ten, five, and one percent levels, respectively.
60
TABLE 6
SHARED AUDITORS AND LENDER LOAN LOSS PROVISION VALIDITY
BY COMMERCIAL LOAN PORTFOLIO

CHGOFF
t+1
= α + β
1
LSHAREAUDOFF
t
+ β
2
LSHAREAUDOFF
t
*LLP
t
+ Controls
t
+ ϵ (1)

(1) (2) (3) (4)
Full Sample PSM Sample
Predicted Commercial Loans Commercial Loans
VARIABLES Sign Low High Low High

LSHAREAUDOFF -0.000 0.000 0.000 0.001
(-0.201) (0.508) (0.202) (0.451)
LSHAREAUDOFF*LLP + 0.192 0.219** 0.124 0.246*
(1.409) (2.182) (0.859) (1.913)
LSPECIALIST -0.002 -0.001 -0.000 -0.001
(-1.503) (-0.855) (-0.186) (-0.900)
LSPECIALIST*LLP + 0.413*** -0.044 0.307 0.062
(2.856) (-0.353) (1.493) (0.410)
LLP + 0.459* 0.389*** 0.870*** 0.145*
(1.948) (5.575) (4.893) (1.794)
LLSIZE + 0.003 0.001 0.006* 0.000
(1.224) (0.913) (1.831) (0.142)
NONACC + 0.034 0.120* -0.155 0.045
(0.208) (1.680) (-1.512) (0.268)
CONSTANT -0.041 -0.014 -0.108* 0.003
(-1.113) (-0.650) (-1.787) (0.067)

Observations 155 251 80 93
Adjusted R-squared 0.809 0.763 0.862 0.768

This table presents the relationship between shared auditors and the validity of loan loss provisions by the lender’s
commercial loan portfolio. Low (high) portfolios have less (greater than or equal to) than twenty percent commercial
loans in their portfolio. Columns (1) and (2) present the analysis using the full sample. Columns (3) and (4) present
the analysis using a propensity-score matched sample. CHGOFF is the next period’s net loan charge-offs scaled by
total assets. LLP is loan-loss provision during year t scaled by beginning total assets. LSHAREAUDOFF equals one
if the lender has greater than the median number of shared auditor office relationships, and zero otherwise.
LSPECIALIST is an indicator variable that equals one if the lender’s auditor has the largest market share of audit
fees in a two-digit SIC category year, and zero otherwise. LLSIZE is the log of total assets at the beginning of year t.
NONACC is non-performing loans at the end of year t scaled by beginning total assets. All continuous variables are
winsorized at the 1% and 99% levels. Year indicators control for year fixed effects. Standard errors are clustered by
company. T-statistics are reported in parentheses. *, **, and *** denote significance at the ten, five, and one percent
levels, respectively.
61
TABLE 7
SHARED AUDITORS AND LENDER AUDIT FEES
BY COMMERCIAL LOAN PORTFOLIO
LAUDITFEE = α + β
1
LSHAREAUDOFF + Controls + ϵ (3)
(1) (2) (3) (4)
Full Sample PSM Sample
Predicted Commercial Loans Commercial Loans
VARIABLES Sign Low High Low High
LSHAREAUDOFF +/- 0.141 0.401*** 0.023 0.412***
(0.873) (3.148) (0.174) (4.728)
LSPECIALIST + 0.211* 0.304*** 0.320** 0.160
(1.763) (2.774) (2.476) (0.940)
LSIZE + 0.700*** 0.550*** 0.694*** 0.349***
(9.643) (7.610) (9.028) (2.946)
LLOSS + -0.002 0.139 -0.079 0.247*
(-0.011) (0.943) (-0.389) (2.010)
SECURITIES + -0.143 -0.048 -0.085 -0.115
(-1.428) (-0.270) (-0.739) (-0.493)
NONPERFORM + 0.587 10.221*** 5.238 20.393***
(0.103) (2.835) (0.423) (3.381)
CHGOFFCY + 2.565 3.108 2.692 -10.812
(0.316) (0.461) (0.294) (-1.628)
COMMLOAN +/- -3.302** -0.533 -4.918*** -2.353
(-2.553) (-0.548) (-3.224) (-1.072)
CONSUMERLOAN +/- 0.242 -2.209* 0.479 -4.528
(0.300) (-1.862) (0.604) (-1.431)
RELOAN +/- 0.478 -2.589** 0.631 -3.662*
(0.750) (-2.629) (1.076) (-1.707)
INTANG + -1.008 6.425 0.510 12.015
(-0.371) (1.255) (0.192) (1.634)
CAPRATIO + 0.018 0.070** 0.056** 0.104**
(0.643) (2.218) (2.127) (2.149)
EXEMPT - -0.350* -0.410*** 0.266 0.150
(-1.983) (-3.089) (1.460) (0.890)
SENSITIVE - 0.618 -0.636* 0.238 -0.741
(1.384) (-1.921) (0.581) (-1.390)
CONSTANT 2.031 5.174** 1.189 8.851**
(1.181) (2.492) (0.653) (2.211)

Observations 157 232 102 153
Adjusted R-squared 0.931 0.884 0.928 0.792
This table presents the relationship between shared auditors and lender audit fees by the lender’s commercial loan portfolio. Low
(high) portfolios have less (greater than or equal to) than twenty percent commercial loans in their portfolio. Columns (1) and (2)
present the full sample. Columns (3) and (4) present a propensity-score matched sample. LAUDITFEE is the natural log of lender
audit fees. LSHAREAUDOFF equals one if the lender has greater than the median number of shared auditor office relationships, and
zero otherwise. LSPECIALIST equals one if the lender’s auditor has the largest market share of audit fees in a two-digit SIC category
year, and zero otherwise. LSIZE is the natural log of total assets. LLOSS equals one if net income is negative, and zero otherwise.
SECURITIES is one less total securities deflated by total assets. NONPERFORM is non-performing loans divided by lagged total
loans. CHGOFFCY is net charge-offs deflated by lagged total assets. COMMLOAN is total commercial and agricultural loans divided
by total loans. CONSUMERLOAN is total consumer loans divided by total loans. RELOAN is total real estate loans divided by total
loans. INTANG is intangible assets divided by total assets. CAPRATIO is total risk-adjusted capital ratio. EXEMPT equals one if a
bank is exempt from FDICIA and Section 404 of SOX, and zero otherwise. SENSITIVE is rate-sensitive assets minus rate-sensitive
liabilities. All continuous variables are winsorized at the 1% and 99% levels. Year indicators control for year fixed effects. Standard
errors are clustered by company. T-statistics are reported in parentheses. *, **, and *** denote significance at the ten, five, and one
percent levels, respectively.
62
TABLE 8
SHARED AUDITORS AND BORROWER GOING CONCERN REPORTING
BY LOAN IMPORTANCE

OPINION = α + β
1
SHAREAUDOFF + Controls + ϵ (2)
(1) (2) (3) (4)
Full Sample Distressed Sample
Predicted Loan Importance Loan Importance
VARIABLES Sign Low High Low High
SHAREAUDOFF +/- 1.827 8.411** 1.607 16.897**
(1.118) (2.228) (1.018) (2.574)
SPECIALIST +/- -0.569 1.129 -0.812 3.464*
(-0.995) (0.978) (-1.221) (1.714)
LSPECIALIST +/- 0.197 -0.160 -0.068 0.551
(0.328) (-0.152) (-0.105) (0.397)
ZMIJEVSKI + 1.184 -6.965* 1.561 -6.383
(0.439) (-1.687) (0.589) (-1.242)
BLLOSS + 0.137 -6.991** 0.000 -14.443**
(0.148) (-2.419) (0.000) (-2.405)
AGE - -0.144 -1.712* -0.185 -4.087**
(-0.322) (-1.688) (-0.380) (-2.136)
TOTLEV +/- -1.928 8.847** -2.666 13.163***
(-0.696) (2.072) (-1.056) (2.608)
CLEV +/- -0.088 -54.195** 1.378 -122.781**
(-0.016) (-2.418) (0.258) (-2.305)
CFO - 3.070 -46.797** 4.289 -93.000**
(0.899) (-2.511) (1.205) (-2.286)
RISKY +/- -1.338 -11.047** -0.640 -27.485
(-0.834) (-1.975) (-0.438) (-0.142)
SIZE - 0.162 0.618 0.087 0.938
(0.564) (1.146) (0.284) (1.338)
ROA - -4.542 -46.061** -1.252 -105.215**
(-1.558) (-2.399) (-0.372) (-2.257)
MB - 0.084 0.411 0.080 0.715
(0.688) (1.616) (0.603) (1.455)
SHAREISSUE - 0.049 -9.130*** -0.403 -20.090***
(0.077) (-2.907) (-0.585) (-2.585)
FUTUREFINANCE - -0.874 -0.188 -1.173 -2.069
(-1.209) (-0.121) (-1.494) (-0.831)
INVESTMENTS +/- 8.667** 68.479*** 13.505** 142.109***
(2.360) (3.043) (2.304) (2.962)
RLAG + 0.525* 2.413*** 0.344** 2.668***
(1.822) (3.476) (2.246) (3.128)
CONSTANT -5.499 -20.288*** -1.983 -18.993*
(-1.404) (-2.662) (-0.646) (-1.694)

Observations 91 108 82 103
Pseudo R-squared 0.198 0.622 0.196 0.722
ROC Area 0.7784 0.9616 0.7841 0.9795

63
This table presents the relationship between shared auditors and going concern reporting by lender loan importance
(less than or greater than or equal to the median loan size) for subsequently bankrupt firms. Columns (1) and (2)
presents the analysis using the full sample. Columns (3) and (4) presents the analysis using a sample of distressed
firms, which are firms with losses in the current year and negative cash flows from operations. OPINION is an
indicator variable equal to one the auditor issued a going concern opinion, and zero otherwise. SHAREAUDOFF is
an indicator variable equal to one if the lender and borrower received audit opinions from the same auditor office,
and zero otherwise. SPECIALIST is an indicator variable that equals one if the borrower’s auditor has the largest
market share of audit fees in a two-digit SIC category year, and zero otherwise. LSPECIALIST is an indicator
variable that equals one if the lender’s auditor has the largest market share of audit fees in a two-digit SIC category
year, and zero otherwise. ZMIJEWSKI is a bankruptcy measure with higher values indicating higher probability of
bankruptcy. BLLOSS is an indicator variable that equals one if the firm had a loss in the prior year, and zero
otherwise. AGE is the log of number of years the company has been publicly traded. TOTLEV is total liabilities over
total assets at the end of the fiscal year. CLEV is the change in leverage during the year. CFO is operating cash flow
scaled by total assets. RISKY is an indicator variable equal to one if the company operated in a risky industry (SIC
codes 2833, 2836, 3570, 3577, 3600, 3674, 7372, 7379, 8731, and 8734). SIZE is the log of total assets. ROA is net
income scaled by average total assets. MB is the market value of equity scaled by book value of equity at the end of
year t. SHAREISSUE is an indicator variable equal to one when a firm issued equity in the fiscal year, and zero
otherwise. FUTUREFINANCE is an indicator variable equal to one when the firm issues equity or long-term debt in
the subsequent year, and zero otherwise. INVESTMENTS are short- and long-term investment securities (including
cash and cash equivalents) scaled by total assets at year-end. RLAG is the number of days between the fiscal year-
end and the opinion date. All continuous variables are winsorized at the 1% and 99% levels. T-statistics are reported
in parentheses. *, **, and *** denote significance at the ten, five, and one percent levels, respectively.

64
TABLE 9
SHARED AUDITORS AND BORROWER AUDIT FEES
BY LOAN IMPORTANCE
BAUDITFEE = α + β
1
SHAREAUDOFF + Controls + ϵ (4)
(1) (2) (3) (4)
Full Sample PSM Sample
Predicted Loan Importance Loan Importance
Variables Sign Low High Low High
SHAREAUDOFF +/- -0.038 0.132*** 0.004 0.113***
(-0.658) (2.729) (0.091) (2.641)
SPECIALIST + 0.034* 0.086*** 0.082** 0.056
(1.707) (3.617) (2.141) (1.572)
LSPECIALIST + 0.031** 0.039*** 0.041 0.105***
(2.439) (2.880) (1.139) (3.158)
SIZE + 0.498*** 0.553*** 0.512*** 0.559***
(53.899) (53.849) (31.719) (39.124)
CURR - -0.078*** -0.074** -0.115*** -0.083**
(-3.004) (-2.043) (-2.950) (-2.093)
QUICK - 0.006 -0.015 0.017 -0.002
(0.229) (-0.353) (0.408) (-0.048)
ROA - -0.428*** -0.544*** -0.438* -0.629***
(-5.383) (-5.148) (-1.866) (-2.787)
LEV + -0.195*** -0.086 -0.450*** -0.241*
(-3.656) (-1.060) (-3.900) (-1.957)
BLOSS + 0.131*** 0.041 0.086 0.011
(5.648) (1.491) (1.472) (0.193)
NUMOPESEGMENTS + 0.027 0.030 0.090** -0.032
(1.309) (1.263) (2.435) (-0.902)
NUMFOREIGNSEGMENTS + 0.221*** 0.200*** 0.262*** 0.160***
(11.178) (7.843) (6.516) (3.595)
AUDITORSWITCH +/- 0.010 -0.023 -0.142 -0.045
(0.344) (-0.599) (-1.346) (-0.538)
OPINION + 0.159*** 0.228** 0.500*** 0.377***
(3.102) (2.088) (2.877) (2.801)
DECYE + 0.051* 0.061* 0.037 0.077
(1.925) (1.677) (0.775) (1.600)
CONSTANT 9.744*** 9.291*** 9.532*** 9.535***
(64.584) (70.189) (36.219) (66.276)

Observations 19,717 19,698 1,555 2,957
Adjusted R-squared 0.777 0.820 0.861 0.881
This table presents the relationship between shared auditors and borrower audit fees based on importance of the loan to the lender (less than or
greater than or equal to the median loan size). Columns (1) and (2) present the analysis using the full sample. Columns (3) and (4) present the
analysis using a propensity-score matched sample. BAUDITFEE is the natural log of borrower audit fees. SHAREAUDOFF is an indicator variable
equal to one if the lender and borrower received audit opinions from the same auditor office, and zero otherwise. SPECIALIST is an indicator
variable that equals one if the borrower’s auditor has the largest market share of audit fees in a two-digit SIC category year, and zero otherwise and
LSPECIALIST is an indicator variable that equals one if the lender’s auditor has the largest market share of audit fees in a two-digit SIC category
year, and zero otherwise. SIZE is the natural log of assets. CURR is the ratio of current assets to current liabilities. QUICK is the ratio of current
assets excluding inventories to current liabilities. ROA is net income scaled by average total assets. LEV is the sum of current and long-term debt
scaled by average total assets. BLOSS is an indicator variable equal to one if the firm has negative net income in the current year, and zero otherwise.
NUMOPESEGMENTS is the natural logarithm of one plus the number of operating segments. NUMFOREIGNSEGMENTS is the natural logarithm
of one plus the number of foreign segments. AUDITORSWITCH is an indicator variable equal to one if the firm switched auditors for the current
year, and zero otherwise. OPINION is an indicator variable equal to one if the firm received a going concern opinion, and zero otherwise. DECYE is
an indicator variable equal to one if the firm has a December fiscal year-end, and zero otherwise. All continuous variables are winsorized at the 1%
and 99% levels. Year and industry indicators based on two-digit SIC codes control for year and industry fixed effects. Standard errors are clustered
by company. T-statistics are reported in parentheses. *, **, and *** denote significance at the ten, five, and one percent levels, respectively.
65

TABLE 10
ALTERNATE MEASURES OF SHARED AUDITORS

Panel A: Lender Loan Loss Provision Validity

CHGOFF
t+1
= α + β
1
LSHAREAUDOFF
t
+ β
2
LSHAREAUDOFF
t
*LLP
t
+ Controls
t
+ ϵ (1)

(1) (2) (3)
VARIABLES Predicted Sign
Continuous
Measure
Materiality
Measure
Proportion
Measure

LSHAREAUDOFF -0.000 -0.001 0.000
(-0.641) (-0.612) (0.043)
LSHAREAUDOFF*LLP + 0.027*** 0.250** 0.007**
(2.945) (2.318) (2.355)
LSPECIALIST -0.002* -0.003** -0.003**
(-1.851) (-2.383) (-2.311)
LSPECIALIST*LLP + 0.207 0.379** 0.326**
(1.626) (2.579) (2.211)
LLP + 0.335*** 0.172 0.326***
(3.633) (1.418) (3.341)
LLSIZE + 0.001 0.001 0.001
(0.786) (0.756) (1.058)
NONACC + 0.166* 0.137 0.164
(1.705) (1.360) (1.635)
CONSTANT -0.010 -0.010 -0.018
(-0.510) (-0.470) (-0.815)

Observations 406 406 406
Adjusted R-squared 0.773 0.771 0.770

This table presents the relationship between shared auditors and the validity of loan loss provisions, which is how well loan loss
provisions map into actual net charge-offs with alternate measures of the lender shared auditor (LSHAREAUDOFF). In column (1),
LSHAREAUDOFF is the number of shared auditor office relationships by lender. In column (2), LSHAREAUDOFF equals one if the
total value of shared auditor loans is greater than or equal to 5% of the lender’s earnings before the loan loss provision, and zero
otherwise. In column (3), LSHAREAUDOFF is the total value of share auditor loans divided by total assets. CHGOFF is the next
period’s net loan charge-offs scaled by total assets. LLP is loan-loss provision during year t scaled by beginning total assets.
LSPECIALIST is an indicator variable that equals one if the lender’s auditor has the largest market share of audit fees in a two-digit
SIC category year, and zero otherwise. LLSIZE is the log of total assets at the beginning of year t. NONACC is non-performing loans
at the end of year t scaled by beginning total assets. All continuous variables are winsorized at the 1% and 99% levels. Year indicators
control for year fixed effects. Standard errors are clustered by company. T-statistics are reported in parentheses. *, **, and *** denote
significance at the ten, five, and one percent levels, respectively.

66
TABLE 10 (continued)
ALTERNATE MEASURES OF SHARED AUDITORS

Panel B: Lender Audit Fees
LAUDITFEE = α + β
1
LSHAREAUDOFF + Controls + ϵ (3)
(1) (2) (3)
VARIABLES Predicted Sign
Continuous
Measure
Materiality
Measure
Proportion
Measure
LSHAREAUDOFF + 0.020** 0.243** 0.005**
(2.210) (2.527) (2.164)
LSPECIALIST + 0.310*** 0.335*** 0.337***
(3.263) (3.523) (3.403)
LSIZE + 0.662*** 0.656*** 0.685***
(11.711) (12.271) (13.287)
LLOSS + 0.136 -0.009 0.129
(0.952) (-0.057) (0.902)
SECURITIES + -0.272** -0.286** -0.263**
(-2.225) (-2.328) (-2.210)
NONPERFORM + 4.249 4.183 5.044
(1.073) (1.049) (1.278)
CHGOFFCY + -1.738 0.810 -0.749
(-0.256) (0.117) (-0.110)
COMMLOAN +/- 0.445 0.408 0.472
(0.529) (0.489) (0.562)
CONSUMERLOAN +/- -0.191 -0.051 -0.151
(-0.234) (-0.065) (-0.188)
RELOAN +/- -0.336 -0.377 -0.374
(-0.511) (-0.577) (-0.572)
INTANG + 0.849 0.548 0.198
(0.311) (0.212) (0.075)
CAPRATIO + 0.062*** 0.061*** 0.061**
(2.662) (2.730) (2.571)
EXEMPT - -0.359*** -0.408*** -0.349***
(-3.128) (-3.437) (-3.006)
SENSITIVE - -0.228 -0.364 -0.297
(-0.674) (-1.061) (-0.885)
CONSTANT 2.081 2.230 1.702
(1.465) (1.568) (1.236)

Observations 389 389 389
Adjusted R-squared 0.895 0.896 0.894

This table presents the relationship between shared auditors and lender audit fees with alternate measures of the lender shared auditor
(LSHAREAUDOFF). In column (1), LSHAREAUDOFF is the number of shared auditor office relationships by lender. In column (2),
LSHAREAUDOFF equals one if the total value of shared auditor loans is greater than or equal to 5% of the lender’s earnings before
the loan loss provision, and zero otherwise. In column (3), LSHAREAUDOFF is the total value of share auditor loans divided by total
assets. LAUDITFEE is the natural log of lender audit fees. LSPECIALIST equals one if the lender’s auditor has the largest market
share of audit fees in a two-digit SIC category year, and zero otherwise. LSIZE is the natural log of total assets. LLOSS equals one if
net income is negative, and zero otherwise. SECURITIES is one less total securities deflated by total assets. NONPERFORM is non-
performing loans divided by lagged total loans. CHGOFFCY is net charge-offs deflated by lagged total assets. COMMLOAN is total
commercial and agricultural loans divided by total loans. CONSUMERLOAN is total consumer loans divided by total loans. RELOAN
is total real estate loans divided by total loans. INTANG is intangible assets divided by total assets. CAPRATIO is total risk-adjusted
capital ratio. EXEMPT equals one if a bank is exempt from FDICIA and Section 404 of SOX, and zero otherwise. SENSITIVE is rate-
sensitive assets minus rate-sensitive liabilities. All continuous variables are winsorized at the 1% and 99% levels. Year indicators
control for year fixed effects. Standard errors are clustered by company. T-statistics are reported in parentheses. *, **, and *** denote
significance at the ten, five, and one percent levels, respectively.
67

TABLE 11
SHARED AUDITORS AND GEOGRAPHY EFFECTS

(1) (2) (3) (4)
VARIABLES
Predicted
Sign
Lender
Audit Quality
Borrower
Audit Quality
Lender
Audit Fee
Borrower
Audit Fee

SHAREAUDOFF + 0.272** 2.260** 0.278** 0.074*
(2.630) (2.304) (2.408) (1.665)

Observations 406 199 389 39,415
Adjusted R-squared 0.771 0.277 0.898 0.821
ROC Area 0.8351

This table reexamines the relationship between shared auditors and the various audit quality measures with an
additional control for the effects of geography, in which an indicator variable equals one if the lender (borrower) is
located in the same city as its auditor. Column (1) presents the results of Model (1). Column (2) presents the results
of Model (2). Column (3) presents the results of Model (3). Column (4) presents the results of Model (4).
SHAREAUDOFF is the shared auditor variable of interest in each of the models. T-statistics are reported in
parentheses. *, **, and *** denote significance at the ten, five, and one percent levels, respectively.

Abstract (if available)

Abstract Auditor competency is a key element in explaining the supply of audit quality, yet our understanding of the drivers of auditor competency in the archival literature is limited. This study uses an archival approach to examine whether sharing auditors among related firms results in information spillovers that improve audit quality. I find that audit quality improves for both borrowers and lenders who share the same auditor office. Specifically, lenders who share an auditor office with their borrowers have more accurate loan loss provisions, especially lenders with larger commercial loan portfolios

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Can I borrow your footnotes? Learning and network benefits of footnote similarity

PDF

Who manages the firm matters: the incremental effect of individual managers on accounting quality

PDF

Share repurchases: how important is market timing?

PDF

Does fair value accounting exacerbate the pro-cyclicality of bank lending?

PDF

Accrual quality and expected returns: the importance of controlling for cash flow shocks

PDF

Gone with the big data: institutional lender demand for private information

PDF

The effects of accounting performance and professional relationships on promotion, dismissal, and transfer decisions in a conglomerate

PDF

Social movements and access to credit

PDF

Effectiveness of the SEC’s comment letters in initial public offerings

PDF

Understanding virality of YouTube video ads: dynamics, drivers, and effects

PDF

Dividend policy, earnings announcements, and analysts' earnings forecasts

PDF

Essays on information, incentives and operational strategies

PDF

Essays in corporate finance

PDF

Essays in tail risks

PDF

That's just what this country needs: another film that's a flop at the flicks: a PR perspective on the success of home-grown films at the Australian box office

PDF

Essays on service systems with matching

PDF

Creating a water quality geodatabase for the West Hawai‘i Island region

PDF

Disclosure of changes in taste: implications for companies and consumers

PDF

The role of the timing of school changes and school quality in the impact of student mobility: evidence from Clark County, Nevada

PDF

Novel and efficient schemes for security and privacy issues in smart grids

Asset Metadata

Creator Ton, Karen (author)

Core Title Do shared auditors improve audit quality? Evidence from banking relationships

School Marshall School of Business

Degree Doctor of Philosophy

Degree Program Business Administration

Publication Date 07/30/2015

Defense Date 05/05/2015

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag audit quality,auditor competencies,banking relationship,OAI-PMH Harvest,shared auditors

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor DeFond, Mark (committee chair), Erkens, David H. (committee member), Murphy, Kevin J. (committee member), Soliman, Mark (committee member), Zhang, Jieying (committee member)

Creator Email kton@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-615665

Unique identifier UC11303540

Identifier etd-TonKaren-3757.pdf (filename),usctheses-c3-615665 (legacy record id)

Legacy Identifier etd-TonKaren-3757.pdf

Dmrecord 615665

Document Type Dissertation

Format application/pdf (imt)

Rights Ton, Karen

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA