API

Volume Three
2003

Audit, Nonaudit, and Information

Technology Fees: Some Empirical Evidence
Lawrence J. Abbott, Susan Parker, Gary F. Peters, and Dasaratha V. Rama

ABSTRACT: This study examines the purchase of nonaudit services by SEC audit cli-
ents. Effective February 5, 2001, the SEC has required the disclosure and description of
audit, nonaudit, and information technology fees paid to the incumbent auditor. The SEC
enacted this rule due to concerns about auditor independence. Our study reviews these
disclosed fees contained in the proxy statements of 2,795 firms since the effective date of
the new SEC disclosure rules. The results indicate that the actual incidence and magnitude
of nonaudit fees is much greater than the information relied upon during the SEC rule-
making deliberations. Consequently, our study calls into question the data used during SEC
rule-making process, as well as the clarity of the current audit and nonaudit service distinc-
tions. Our results also suggest that the nonaudit service landscape has changed dramati-
cally since the last period of publicly available fee data (1978–1981). Our results provide at
least two public policy implications. First, the SEC should maintain, at a minimum, the cur-
rent disclosure environment. Second, the SEC should strive to improve the clarity and granu-
larity of the disclosures. Our results also indicate a clear need for additional research regard-
ing the provision of nonaudit services.

INTRODUCTION
The objective of this study is to examine the magnitude and frequency of nonaudit services
(NAS) purchased by SEC audit clients from their incumbent auditors. This issue, which has contin-
ued to be an area of concern for regulators and auditors alike, addresses an intangible, but vital
aspect of the accounting profession: the investing public’s confidence in the independence of the
external auditor. The growth in NAS and the concomitant independence-related concerns moti-
vated the SEC to originally propose rules that would have prohibited audit firms from providing
certain NAS to their audit clients. The AICPA strenuously opposed such measures and claimed the
proposed rules would hurt the public interest and the auditing profession. Public hearings concern-
ing these rules commenced in July 2000.
During these hearings, the SEC (2000a, 2000b) relied upon two important suppositions during
the rule-making process: one relating to the proportion of audit clients purchasing NAS from their

Lawrence J. Abbott is an Assistant Professor at the University of Memphis, Susan Parker is an Assistant Professor at
Santa Clara University, Gary F. Peters is an Assistant Professor at the University of Arkansas, and Dasaratha V. Rama
is a Professor at Florida International University.

We thank two anonymous reviewers, Greg Trompeter, John Eichenseher, Pamela Roush, K. Raghunandan, and partici-
pants at the 2002 Midyear American Accounting Association Auditing Section Meeting for their helpful comments and
encouragement. All data are from public available sources. All errors are ours.
Abbott, Parker, Peters, and Rama 2

auditors and the other relating to the magnitude of NAS fees. Both suppositions were based upon
data gathered from annual reports filed by the public accounting firms with the SEC Practice Sec-
tion (SECPS) of the AICPA. The debate over the proposed rules became heated and appeared
headed toward the courts. Nearing the end of his tenure, then SEC Chairman Levitt agreed to a
compromise in November 2000. Instead of prohibiting certain NAS, the SEC took a “market-based”
approach whereby SEC registrants had to disclose in their annual proxy statements amounts paid
to their incumbent auditors for audit fees, financial information systems design, and implementa-
tion services (FIS) fees, and “other” nonaudit (non-FIS) fees.1 In the end, the SEC argued that,
based upon the data it had gathered from the SECPS reports, the disclosures would not seriously
harm the profession, and that the benefit to the financial community would outweigh any costs.
In this paper, we examine the magnitude and frequency of FIS and other nonaudit fees in
relation to audit fees. We use data provided by a sample of 2,795 firms that filed their proxy state-
ments between February 5, 2001 and June 30, 2001. We then calculate the ratio of the total NAS
fees (FIS fees plus other nonaudit services fees) to audit fees (henceforth, the NAS fee ratio) for
these sample firms. We find a mean (median) NAS fee ratio of audit fees of 1.60 (0.96). Our 75th
and 25th percentiles for the NAS fee ratio were 1.969 and 0.402, respectively. Moreover, the maxi-
mum ratio was 45.56—implying an auditor received forty-five times as much compensation related
to NAS than audit services. Our empirical evidence is dramatically different from SECPS data
relied upon by the SEC during its rule making. Our results suggest that the SEC may have under-
estimated the pervasiveness of NAS purchases among its registrants. Most importantly, the data
we do have, from these first proxies, is so vastly different from what was widely anticipated, that the
entire area of nonaudit purchases by SEC registrants from their audit firms is now much more
important for policy makers and a much more fertile field for researchers.
Perhaps not surprisingly, our findings also support the contention that the audit and nonaudit
fee landscape has changed remarkably since the previous disclosure regime of ASR No. 250 of
1978–1981. Univariate tests document a very significant, positive relation between client size and
the NAS fee ratio. We also find significant univariate differences in the NAS fee ratio between Big
5 and non-Big 5 auditors, as well as significant intra-Big 5 differences in the NAS fee ratio. Multivari-
ate tests generally confirm our univariate findings.
The evidence provided herein evokes at least two policy implications. First and foremost, we
believe that it is in the public interest for the SEC to maintain, at a minimum, the current disclosure
environment. The ultimate test of the impact of NAS on audit quality is to examine the impact of
NAS on audit failures. However, a powerful test of this proposition requires numerous observa-
tions, mandating that the disclosures be maintained for a relatively long period of time. It should be
noted that while the recently enacted Sarbanes-Oxley Act proscribes certain NAS, it does not
outright ban all NAS. Moreover, given the relatively large magnitudes and frequency of nonaudit
service purchases documented in this paper, it remains an important empirical question whether
the Sarbanes-Oxley Act will impact the purchase of NAS and, in turn, positively impact the audit
function. Our study provides a benchmark reference for future studies on this issue.
A second policy implication is that a more detailed disclosure of NAS types would likely assist
market participants in determining the impact of NAS on auditor independence. To wit, academic

1
For expositional purposes, the term “other” nonaudit fees is used in this paper to describe all non-FIS, nonaudit fees under the
most recent disclosure environment (i.e., after February 5, 2001). The term NAS is used in this paper to refer to both “other”
nonaudit fees and FIS fees.

Accounting and the Public Interest Volume 3, 2003

Abbott, Parker, Peters, and Rama 3

research suggests that market participants view recurring versus nonrecurring consulting engage-
ments differently (Swanger and Chewning 2001; Lowe et al. 1999). However, unlike ASR No. 250,
the current disclosures do not differentiate between recurring and nonrecurring services. Further-
more, in an effort to create a resolution to the debate in a timely fashion, the SEC’s new disclosure
rules likely sacrificed the clarity of these fee disclosures. In particular, as opposed to the disclo-
sures under ASR No. 250, the current disclosures require grouping NAS fees into two broad nonaudit
fee categories. Consequently, many NAS fees that auditors and others would consider “audit-
related” (such as the audit of a company’s pension and benefit plans) are now considered part of
the larger umbrella of nonaudit services.2 This may actually obfuscate the total audit relationship
between client and auditor from the vantage point of investors. This is epitomized by Liesman et al.
(2002) who state:
As for auditors, last year marked the first time that companies were required to disclose
the amount of fees they paid their independent accounting firms for nonaudit services. Yet
the required descriptions of those services remain vague at best.
The remainder of this paper is organized as follows. The next section discusses the historical
background of ASR No. 250 (and its subsequent repeal), followed by an account of the July 2000
hearings and a description of the final rules. This is succeeded by a sample selection and results
section and a discussion of current developments. The paper ends with a conclusion that dis-
cusses our policy implications in greater detail and offers suggestions for future research on some
of the issues raised in the paper.

BACKGROUND
Previous U.S. Disclosure Regimes and Concerns
about the Impact of NAS on Auditor Independence

A vital attribute of the external audit is the independence of the auditor. This requirement
entails, among other things, that the auditor maintains independence from client management.
Because of management’s influence in the decision process regarding the hiring of and subse-
quent reappointment of the auditor, shareholders have to be concerned about the independence of
the auditor in safeguarding investors’ interests. As actual independence is impossible to observe,
shareholders have to rely on the auditor’s reputation, professional governing body, and regulatory
oversight and various “signals” of independence. One such signal may be the degree of perceived
economic bonding between the auditor and client management.
The provision of NAS to audit clients has long been theorized to increase the economic bond
between the auditor and client (Roberts 2001). This bond could lead to the perception of impaired
auditor independence because: (1) the audit firm may be unwilling to criticize the work done by its
consultancy divisions, and (2) the audit firm does not want to lose lucrative consultancy services
provided to the audit client and, therefore, may be more reluctant to disagree with management’s
interpretations of accounting matters (Beck et al. 1988; Pany and Reckers 1983, 1984). Along
these lines, ASR No. 250 was enacted on June 29, 1978 under the Chairmanship of Harold Williams,

2
One Big 5 auditor (Ernst & Young) recommends its audit clients to disclose an additional category called Audit-Related Fees in
order to more accurately depict the “total audit relationship” (Ernst & Young 2001).

Accounting and the Public Interest Volume 3, 2003

Abbott, Parker, Peters, and Rama 4

an activist, Democratic appointee under the Carter Administration.3 The reasoning behind ASR No.
250 was for shareholders to determine the degree to which clients purchased NAS from their
incumbent auditors and any independence related issues that may arise thereof.
Similar to the recent SEC disclosure requirements, ASR No. 250 was extremely contentious.
The ASR No. 250 disclosure arguments revolving around the provision of nonaudit services are
remarkably similar to those offered during the recent SEC disclosure hearings. For example, in
justifying the provision of nonaudit services Arthur Young & Company (1978) asserted advisory and
analytical services were a “natural outgrowth of the auditing practice.” As “businesses (become)
more complex, tax laws more intricate, and computerization a significant factor in accounting pro-
cesses, companies (request) consultation and advice in a variety of areas from their independent
accountants” (Arthur Young & Company 1978). Clients refer to their incumbent auditors due to the
“knowledge the auditor gains, during the course of the audit, of the many facets of a company’s
activities, including its operations, objectives, financial affairs, personnel, production methods and
so on” (Arthur Young & Company 1978).
In contrast, the auditor independence concerns were best reflected in a letter to the (then)
Chairman of the SEC Practice Section by the Chairman of the Public Oversight Board:
The POB believes that there is possibility of damage to the profession and the users of the
profession’s services in an uncontrolled expansion of management advisory services to
audit clients. Investors and others need a public accounting profession that performs its
primary function of auditing financial statements with both the fact and the appearance of
competence and independence. Developments which detract from this will surely dam-
age the professional status of CPA firms and lead to suspicion and doubts that will be
detrimental to the continued reliance of the public upon the profession without further and
more drastic government intrusion. (McCloy 1979)
Among the provisions of ASR No. 250 was the disclosure of audit and several different types
of NAS, among them management planning and organizational control consulting; financial and
accounting systems and controls implementation; tax and actuarial services; and audits of em-
ployee benefit plans. The threshold for disclosure was any NAS greater than 3 percent of audit
fees, either individually or in the aggregate. The ASR No. 250 requirements were very detailed in
that they required registrants to specifically describe each service provided by their incumbent
auditor. Broad categories such as “tax matters” or “management advisory services” were not suffi-
ciently specific.
Research addressing the impact of ASR No. 250 investigated whether the disclosures re-
quired under ASR No. 250 were used by shareholders. Parkash and Venable (1993) found a nega-
tive relation between various agency cost variables and the ratio of NAS to audit fees, suggesting
that client managers were aware of the impact of NAS on shareholders’ perception of auditor
independence. In contrast, Glezen and Millar (1985) generally did not find that shareholder auditor
ratification percentages were negatively related to the NAS to audit fee ratio. Parkash and Venable
(1993) and Glezen and Millar (1985) both noted a relatively low ratio of NAS to audit fee ratios of
0.24 and 0.30, respectively.
In September 1981, ASR No. 250 was repealed by the SEC under the Chairmanship of John
Shad, a market-oriented Republican appointee under the Reagan Administration. The rationale for
3
We are deeply indebted to an anonymous reviewer for suggesting this historical context.

Accounting and the Public Interest Volume 3, 2003

Abbott, Parker, Peters, and Rama 5

the repeal was threefold. First, studies of the disclosures concluded that the auditors were not
providing extensive nonaudit services to their clients (SEC 2000a). Second, the disclosures may
have caused an artificial and detrimental curtailment of NAS (SEC 1982). These two issues were
emphasized by then SEC Chairman Shad—who also noted that marketplace participants seemed
to have little interest in the data provided by the disclosures (Shad 1981). Last, the SEC asserted
that sufficiently useful data would be provided to the SEC practice section of the AICPA (SECPS).
Chairman Shad (1981) stated that the AICPA’s “peer review and other membership requirements”
should afford investors “the requisite degree of assurance that its members consistently conduct
their … practices in accordance with highest professional standards.”
The provision of NAS remained a relatively small issue until the early 1990s, when consulting
work among the Big 6 grew explosively, from 31 percent of the industry’s fees in 1993 to 51 percent
in 1999 (Brynes et al. 2002). It was also at this time that Arthur Levitt became SEC Chairman. Levitt
began voicing his concerns about the magnitude and frequency of NAS engagements early in his
tenure. (See, for example, Levitt 1996a, 1998, 1999, 2000). He criticized audit firms for using
auditing as “a loss-leader retained as a foot in the door for higher-fee consulting services” (Levitt
1996b). Chairman Levitt and the SEC contended that:
• audits were priced low, in the hopes of obtaining the more profitable nonaudit services
• such low-balling led to reduced audit quality, and
• lucrative nonaudit services exacerbated problems with auditor independence.
Levitt (2000) continued to voice his fears that consulting and other services “shorten the dis-
tance between the auditor and management” and that “independence—if not in fact, then certainly
in appearance—becomes a more elusive proposition.” Levitt originally proposed banning all NAS
(including FIS) consulting engagements.
The AICPA and Big 5 auditors strenuously objected to these proposals and stated the rules
could hurt both the profession and public interest. Many auditors and others noted that there are
benefits from the joint supply of audit and nonaudit services. Elliott (SEC 2000d) noted the econo-
mies of scale accruing to the audit client as a result of providing such nonaudit services, as well as
a potential increase in total audit quality due to knowledge spillovers obtained during the provision
of such services. Elliott (SEC 2000d) stated that “improving information systems is complementary
to the audit, which is the improvement of one type of information. When you look at it in that way,
the majority and, in fact, the vast majority of the work that’s done by the accounting firms is actually
in the service of information integrity … which is ultimately to the shareholders’ benefit and the
economy’s benefit.”
In terms of the impact of NAS fees on auditor independence, many practitioners noted the lack
of any empirical studies linking NAS fees to auditor independence. Melancon (SEC 2000d) noted,
“None of the studies reported any conclusive evidence of diminished audit quality or harm to the
public interest as a consequence of public accounting firms providing advisory or consulting ser-
vices to their audit clients.” Melancon (SEC 2000d) criticized the proposed rules and stated, “We
are sincerely concerned, if this rule takes effect, about the profession’s ability to meet the public
interest in the long term.” Melancon also roused the support of the State Societies of CPAs, to
argue against the SEC’s proposals, on the basis that the rules would have a spillover effect on the
smaller CPA firms, who might find themselves held to those same rules, even for their non-SEC
clients. The Big 5 accounting firms also made large political contributions to key legislators such as
Phil Gramm (Republican member of the Ways and Means Committee—the committee that sets the

Accounting and the Public Interest Volume 3, 2003

Abbott, Parker, Peters, and Rama 6

SEC’s annual budget) and Billy Tauzin (Republican Chairman of the House Energy and Commerce
Committee) in order to increase the political pressure on Chairman Levitt.4
The debate became increasingly divisive and appeared headed toward the courts for a long,
drawn-out resolution. Perhaps knowing his tenure was coming to a close, Chairman Levitt began
inviting dialogue from interested parties on the NAS issue in July 2000.

SEC Rule Proposal in July 2000

In response to growing concerns about the effect of NAS fees on auditor independence, the
SEC issued a rule proposal that would have severely restricted the supply of NAS by incumbent
audit firms to public audit clients. In justifying the rule proposal, the SEC (2000a) cited and relied
upon the following representations:
• For the five largest public accounting firms, management advisory services (MAS) fees re-
ceived from audit clients amounted to 10 percent of all revenues in 1999. Almost three-
fourths of audit clients did not purchase any MAS from their auditors in 1999. This means
that purchases of MAS services by one-fourth of firms’ audit clients account for 10 percent of
all firm revenues.
• The proportion of Big 5 audit clients that paid MAS fees in excess of audit fees did not
exceed 1.5 percent until 1997. In 1999, 4.6 percent of Big 5 audit clients paid MAS fees in
excess of audit fees, an increase of over 200 percent in two years.
The SEC gathered the above data from annual reports filed by the public accounting firms with
the SEC Practice Section (SECPS) of the AICPA. The SECPS requires member firms to classify
the number of their audit clients for whom the management consulting fees ratio falls in the follow-
ing categories: 0, 0.01 to 0.25, 0.26 to 0.50, 0.51 to 1.00, and more than 1.00.
It is worthwhile to note that the SECPS used a category titled “Accounting and Auditing,” which
is more expansive than the “Audit Fees” category used by the new SEC rule and ASR No. 250.5
The SECPS required detailed data (such as the proportion of clients with fees exceeding audit
fees) only for management consulting services, otherwise known as management advisory ser-
vices—but not for overall nonaudit services.6

Nonaudit Services (NAS) and Management Advisory/Consulting Services (MAS)

While the SEC expressed its concerns about nonaudit services (NAS), the above SECPS data
cited by the SEC refer only to Management Advisory Services (MAS). MAS is only a subset of NAS

4
In the February 4, 2002 issue of BusinessWeek, Tauzin admits “it’s kind of hard to say there isn’t evidence of a problem …
Arthur Levitt was right” (Carney and Cohn 2002).
5
It is likely that the definitional differences between the SECPS data and the new disclosure rules contributed to the vast
discrepancy between what the SEC relied upon and the data we report in later sections. Furthermore, there was likely interpre-
tational differences existing at the time of the hearings as well. We thank an anonymous reviewer for this comment. Nonethe-
less, it appears that the SECPS data was the sole source of data that the SEC relied upon during the rule-making process.
6
ASR No. 250, which required disclosure of nonaudit fee ratios, was repealed in 1982. One of the reasons offered at the time of
repeal was that similar data would be available with the SECPS. However, the SECPS’s definitions for the categories of fees
were different from the definitions used by the SEC under ASR No. 250 (and under the new rules). Note that there are two other
differences between the SEC’s required disclosure about nonaudit fees and the data available from reports filed with the
SECPS. First, the SEC data are at the company level while the SECPS data are at the audit firm level. Second, data filed with
the SEC are filed under a different legal environment, and are subject to legal and/or monetary penalties.

Accounting and the Public Interest Volume 3, 2003

Abbott, Parker, Peters, and Rama 7

because NAS includes many other services, such as internal audit outsourcing, litigation support
services, audits of benefit plans, and tax consulting services.7
This confusion between NAS and MAS can be observed in the initial SEC rule proposal, at
many of the hearings subsequent to the rule proposal, and in the final SEC rule. Here are some
examples (note: below and elsewhere in this paper, we have added the emphasis).
From the Rule Proposal (SEC 2000a):
Currently, the five largest public accounting firms audit approximately 12,800 public com-
panies. Other public accounting firms audit approximately 3,900 public companies. Ac-
cording to reports filed with the AICPA, of the 12,800 public companies audited by the so-
called “Big 5,” approximately 9,500 did not purchase any consulting services from their
auditor in the most recently reported year. Of the approximately 3,900 registrants that are
audited by other public accounting firms, approximately 3,100 did not purchase any con-
sulting services from their auditor.
For the 12,600 registrants that did not purchase any consulting services from their audi-
tor, the proposed amendments would not have affected their purchase of nonaudit ser-
vices in the most recently reported year … .
Of the approximately 4,100 (3,300 Big 5-audited clients; 800 non-Big 5-audited clients)
registrants that were reported to have purchased nonaudit services from their auditor,
many may have purchased nonaudit services that are not covered by the proposals.
Note that the first paragraph above refers to “consulting services”; the next paragraph mixes
“consulting services” and “nonaudit services”; the third paragraph, even while relying on the data
from the first paragraph, uses the phrase “nonaudit services.”
The confusion persisted during the hearings on the rule proposal. Here are some comments
from the hearings:
Given that a small fraction of audit clients use nonaudit services, I believe the release
points to 25 percent …” (Robert Elliott, Chairman of the AICPA [SEC 2000d])
The point has been made that since only 25 percent of audit clients currently receive
nonaudit services from their audit firm … The fact is that while we perform consulting or
advisory work for a minority of our audit clients in any one year, we use these competen-
cies to help us do a quality audit on almost all of our audit clients every year. (James
Copeland, CEO of Deloitte & Touche [SEC 2000c])
Before issuing a new rule, the SEC was required to perform a cost-benefit analysis of the rule.
As part of the cost-benefit analysis, the SEC (2000b) stated:
To obtain an estimate of the number of individuals and businesses that may benefit, we
note that, in any given year, approximately 74.3 percent of companies purchase only
auditing services from their Big 5 auditor. 8 It may be reasonable, therefore, to estimate
that only twenty-five percent of audit clients will be directly affected by the rule.
7
“Management advisory services (MAS) are consulting activities that may involve providing advice and assistance concerning
an entity’s organization, personnel, finances, operations, systems, or other activities … the MAS department may be divided
further into areas of specialization such as small-business consulting, management information systems development, litiga-
tion support services and actuarial/pension services” (Messier 2000).
8
Footnote 583, inserted here, noted “while we recognize that the set of firms that may purchase such [nonaudit] services may
change from year to year, we have received no evidence to suggest that the fraction of companies that may actually purchase
such services in any given year is different from our estimate.”

Accounting and the Public Interest Volume 3, 2003

Abbott, Parker, Peters, and Rama 8

The SEC did request comments on the accuracy of the data. Specifically, SEC (2000a) noted
that “we also request comment on the accuracy of the estimated number of issuers [SEC regis-
trants] that would be affected by the proposed amendment.” However, the Final Rule (SEC 2000b)
stated:
In the Proposing Release, we stated the burden would fall primarily on one-quarter of
registrants because only one-quarter of registrants receive nonaudit services from their
accountants in any given year. Some commenters disagreed. While it may be true, these
commenters suggested, that only twenty-five percent of registrants receive nonaudit ser-
vices in any given year, a larger percentage receives nonaudit services in some years and
not others.
Footnote 212 of the final rule, which was inserted at the end of the paragraph above, noted:
Because we believed that it would have been useful to have additional data concerning
the revenue mix of accounting firms, as well as the extent to which fees to audit clients for
nonaudit services exceed fees for audits, we solicited comment on revenue data. In addi-
tion, SEC Commissioner Isaac C. Hunt, Jr. informed the Big 5 firms that these data would
help the Commission in its deliberations … However, no data were submitted by any of
the five firms.”
Three of the Big 5 firms (Arthur Andersen, Deloitte & Touche, and KPMG) and the AICPA were
strenuously opposed to the SEC’s actions related to NAS, while the chairman of the SEC had
expressed his strong preferences for some actions that would curtail the supply of NAS to public
audit clients. Viewed in this light, pointing out that a much higher proportion of audit clients pur-
chased NAS could have been counterproductive for the AICPA or the Big 5.

Final SEC Rules

After a contentious debate with the AICPA and some accounting firms, the SEC finally acted to
restrict the supply of some NAS by incumbent auditors to their SEC audit clients. The SEC adopted
a two-pronged approach in response to concerns about the impact of NAS on auditor indepen-
dence. First, it proscribed certain types of nonaudit services. With one exception (outsourcing of
internal audit services), the list of proscribed services mirrors the existing requirements under the
SECPS rules for member firms. In the end, the SEC took a “market-based” approach and sought to
allow marketplace participants to determine the true impact of NAS fees on auditor independence
(Unger 2001). Second, the SEC required registrants to disclose publicly the amounts paid to in-
cumbent auditors for audit, FIS, and other nonaudit services in their annual proxy statements.
Chairman Levitt (SEC 2000c) noted that the disclosure about audit and NAS fees “could help an
investor establish whether (accounting and auditing) judgment calls may lean more toward the
company’s interest or the shareholders’ interests.”
Given the above discussion, we believe it is useful to examine the proportion of SEC regis-
trants that purchase FIS and other NAS from their auditors and the relative magnitude of those
purchases. Our focus is based upon the assertion that financial statement users will find this infor-
mation useful in forming assessments about an auditor’s independence. For example, Robert Ryan,
the Chief Financial Officer of Medtronic Inc., noted:

Accounting and the Public Interest Volume 3, 2003

Abbott, Parker, Peters, and Rama 9

An important consideration in evaluating independence is the magnitude of the fee of the

nonaudit service provided both in absolute dollars and compared to the audit fee … It’s
fair to say that the larger the fee for the consulting service, the more likely independence,
actual or perceived, will be impaired. (SEC 2000c)

DATA AND RESULTS

The new SEC rules about the disclosure of audit, FIS, and other nonaudit fees became effec-
tive for proxies filed on or after February 5, 2001. We obtained fee information from proxies filed by
nonfinancial firms with the SEC between February 5, 2001 and June 30, 2001. We obtained finan-
cial and auditor information from Compustat. From our total sample of 3,241, we lost 446 firms that
did not have Compustat information on total assets, resulting in a sample size for our size-related
tables of 2,795. We lost 317 more firms due to lack of auditor information, yielding a sample of
2,478 for our auditor-related tables.

Sample Description

As shown in Table 1, sample firms exhibited a mean (median) size of $4,421 million ($320
million) in total assets, with a range of $0.2 to $902,210 million. Mean (median) audit fees are
$561,569 ($191,000), mean (median) FIS fees are $207,605 ($0), and mean (median) other fees
are $1,154,352 ($189,773). The maximum audit, FIS, and other nonaudit fees in the sample are
$48,000,000, $46,800,000, and $77,000,000, respectively. The maximum NAS fee ratio is 45.56.
Table 2 shows the fee and ratio data, broken down by size. All fees and fee ratios are increas-
ing in company size.9 The increase in the NAS fee ratio with company size suggests that the
monetary importance of NAS fees to audit firms is much higher than that suggested by the overall
sample medians. Thus, the overall mean (median) NAS fee ratio in Table 1 for the sample of 2,795
companies is 1.60 (0.96). However, total audit fees paid by our sample companies were $1.6
billion, whereas total nonaudit fees paid by those same companies equaled $3.7 billion. Thus, the
overall NAS fee ratio when examining the total dollar amounts is 2.30 ($3.7/$1.6).
Table 3 shows descriptive data, by auditor. In terms of the total NAS fee ratios, the firms rank
as follows: PricewaterhouseCoopers, Andersen, Ernst & Young, KPMG, Deloitte & Touche. All Big
5 firms have significantly higher ratios than non-Big 5 firms.10

Magnitude of NAS fees

As seen in Panel A of Table 4, 1,429 firms (51 percent) reported that the magnitude of NAS
paid to the auditor was greater than the audit fee. When we partitioned the sample by type of
auditor, as seen in Panel B of Table 4, the percentage of clients with NAS fee ratio over 1.00 ranges
from 58.4 percent (PWC) to 15 percent (non-Big 5 firms). In terms of individual audit firms, the
percentage of clients with an NAS fee ratio over 3.00 ranges from 19.5 (PWC) to 1.7 (non-Big 5).

9
Parametric, two-sample t-tests, as well as nonparametric Mann-Whitney tests, indicate significant fee and fee ratio differences
between all size quartiles at p < .05.
10
Parametric, two-sample t-tests, as well as nonparametric Mann-Whitney tests, indicate significant fee ratio differences be-
tween Big 5 and non-Big 5 auditors at p < .05.

Accounting and the Public Interest Volume 3, 2003

Accounting and the Public Interest

Abbott, Parker, Peters, and Rama

TABLE 1
Univariate Statistics: Fees and Ratios
Number of Firms = 2,795

FEES
Variable Mean Median Std. Dev. 25th Quartile 75th Quartile Minimum Maximum
Total Assets (millions) 4421.48 320.290 28,238 79.86 1381.01 0.200 902,210
Audit Fees ($) 561,569 191,000 1,784,086 101,110 425,075 5,000 48,000,000
FIS Fees ($) 207,605 0 1,949,176 0 0 0 46,800,000
Other Fees ($) 1,154,352 189,773 4,522,930 50,000 641,000 0 77,000,000

RATIOS
Variable Mean Median Std. Dev. 25th Quartile 75th Quartile Minimum Maximum
FIS Fees / Audit Fees 0.151 0.000 1.252 0.000 0.000 0.00 44.30
Total NAS Fees / 1.600 0.960 2.233 0.402 1.969 0.00 45.56
Audit Fees
Total NAS Fees = FIS Fees + Other Fees.
Volume 3, 2003

10
Abbott, Parker, Peters, and Rama 11

TABLE 2
Fees and Ratios by Client Size Quartiles
Panel A: Fees
Client Client Assets
Size and Auditors’
Quartile Fees Mean Median Std. Dev. Min. Max.
< $79.86 Assets (millions) 33.079 30.01 21.73 0.2 79.86
(n = 699) Audit Fees 108,621 90,800 82,400 5000 21,295,000
FIS Fees 1,243 0 25,698 0 675,000
Other Fees 83,345 38,000 158,734 0 3,040,000
$79.86 to Assets (millions) 176.12 164.03 68.27 79.95 317.95
$317.95 Audit Fees 190,808 154,160 137,044 15,300 998,800
(n = 698) FIS Fees 15,273 0 111,815 0 1,549,000
Other Fees 224,046 139,450 263,819 0 2,854,000
$317.95 to Assets (millions) 706.114 649.49 288.04 320.29 1,381.01
$1381 Audit Fees 331,543 250,000 318,756 23,237 3,198,500
(n = 699) FIS Fees 49,918 0 395,937 0 7,000,000
Other Fees 539,568 250,000 1,167,707 0 23,252,000
> $1381.01 Assets (millions) 16,764.52 3,989 54,663 1,381 902,210
(n = 699) Audit Fees 1,614,774 720,500 3,261,483 34,000 48,000,000
FIS Fees 732,392 0 3,799,201 0 46,800,000
Other Fees 3,597,575 1,195,000 7,860,193 0 77,000,000

Panel B: Ratios
Size Quartile Fee Ratios Mean Median Std. Dev. Min. Max.
< $79.86 FIS fee ratio 0.012 0.000 0.256 0 6.75
(n = 699) NAS fee ratio 0.730 0.450 0.925 0 7.67
$79.86 to 317.95 FIS fee ratio 0.099 0.000 0.793 0 16.90
(n = 698) NAS fee ratio 1.353 0.900 1.612 0 20.51
$317.95 to $1381 FIS fee ratio 0.162 0.000 1.807 0 44.30
(n = 699) NAS fee ratio 1.741 1.111 2.451 0 45.56
> $1381.01 FIS fee ratio 0.330 0.000 1.520 0 24.19
(n = 699) NAS fee ratio 2.521 1.617 2.783 0 28.85
Size quartiles is based on Total Assets ($ millions).
FIS Fee Ratio = FIS Fees/Audit Fees.
NAS Fee Ratio = (FIS Fees + Other Fees) / Audit Fees.

Accounting and the Public Interest Volume 3, 2003

Abbott, Parker, Peters, and Rama 12

This is in marked contrast to the SECPS data relied upon during the rule-making process, which
suggested that less than 5 percent of the Big 5 audit firms’ clients had nonaudit fees in excess of
the audit fee.

Frequency of NAS Purchases from Incumbent Auditors

Per Table 4, Panel B, only 108 of the 2,478 companies with auditor information (4.4 percent) in
our sample reported that they did not purchase any NAS from their incumbent auditor. When we
partitioned by auditor type, the results were as follows:
• Forty-one of the 241 companies (17 percent) audited by non-Big 5 audit firms did not pur-
chase any NAS from their incumbent auditor.

TABLE 3
Univariate Statistics: Fees and Ratios by Auditor

Panel A: Univariate Statistics—Auditors and Fees

Client Assets and
Firm Auditors’ Fees Mean Median Std. Dev. Min. Max.
Andersen Assets (millions) 2,585 322 10,287 2 171,532
(n = 471) Audit Fees 676,355 214,100 2,626,092 29,500 48,000,000
FIS Fees 241,072 0 2,176,558 0 37,600,000
Other Fees 1,064,054 232,500 2,794,651 0 31,000,000
Deloitte & Assets (millions) 6,920 305 38,440 2 426,794
Touche Audit Fees 783,544 230,000 2,142,899 5,000 20,115,000
(n = 348) FIS Fees 281,459 0 2,373,569 0 41,300,000
Other Fees 1,512,425 207,879 6,484,696 0 77,000,000
Ernst & Assets (millions) 3,155 278 14,953 1 224,720
Young Audit Fees 470,519 207,000 803,930 17,500 6,000,000
(n = 492) FIS Fees 113,851 0 1,274,410 0 23,300,000
Other Fees 1,063,660 228,119 3,376,350 0 51,400,000
KPMG Assets (millions) 5,529 267 53,569 2 902,210
(n = 353) Audit Fees 600,102 190,000 2,029,668 23,000 26,130,000
FIS Fees 238,605 0 2,159,535 0 35,500,000
Other Fees 1,056,730 166,000 4,574,410 0 68,200,000
PWC Assets (millions) 4,051 395 19,761 2 289,760
(n = 573) Audit Fees 707,879 240,700 1,730,898 18,000 18,300,000
FIS Fees 347,989 0 2,573,785 0 46,800,000
Other Fees 1,590,366 320,000 5,145,462 0 68,500,000
Non-Big 5 Assets (millions) 130 26 314 0.02 2,460
(n = 241) Audit Fees 135,738 80,458 240,430 14,000 3,144,822
FIS Fees 660 0 3,892 0 30,000
Other Fees 88,629 25,000 316,804 0 4,348,664
(continued on next page)

Accounting and the Public Interest Volume 3, 2003

Abbott, Parker, Peters, and Rama 13

TABLE 3 (continued)
Univariate Statistics: Fees and Ratios by Auditor

Panel B: Univariate Statistics—Auditors and Fee Ratios

Firm Fee Ratios Mean Median Std. Dev. Min. Max.
Andersen FIS fee ratio 0.327 0.000 2.561 0.000 44.300
(n = 471) NAS fee ratio 1.915 1.151 3.132 0.000 45.560
DT FIS fee ratio 0.138 0.000 0.643 0.000 6.258
(n = 348) NAS fee ratio 1.489 0.954 1.806 0.000 13.490
EY FIS fee ratio 0.079 0.000 0.594 0.000 0.280
(n = 492) NAS fee ratio 1.616 1.126 1.994 0.000 25.520
KPMG FIS fee ratio 0.172 0.000 0.970 0.000 9.781
(n = 353) NAS fee ratio 1.516 0.887 1.971 0.000 15.974
PWC FIS fee ratio 0.189 0.000 1.061 0.000 10.581
(n = 573) NAS fee ratio 1.922 1.305 2.132 0.000 16.045
Non-Big 5 FIS fee ratio 0.006 0.000 0.036 0.000 0.293
(n = 241) NAS fee ratio 0.525 0.323 0.705 0.000 5.002
Sample size equals 2,478 (2,795 less 317 firms without auditor information per Compustat).
FIS Fee Ratio = FIS Fees/Audit Fees.
NAS Fee Ratio = (FIS Fees + Other Fees)/Audit Fees.

• Sixty-seven of the 2,237 companies (2.99 percent) audited by the Big 5 audit firms indicated
that they did not purchase any NAS from their incumbent auditor.11
Recall that throughout the rule-making process, the SECPS data suggested that about 75
percent of the Big 5 audit firms’ clients did not purchase nonaudit services from their auditor.

Multivariate Analysis

In order to more fully understand the impact of size and auditor type on the NAS fee ratio, we
examine the relation between these variables in a multivariate setting. Palmrose (1986) and Barkess
and Simnett (1994) found that larger companies were more likely to purchase nonaudit services,
owing to more complex systems and a wider range of activities. Given that size may proxy for client
complexity and the incumbent demand for systems-related consultancy (Firth 1997), we expect a
positive relation between client size and the NAS fee ratio. We also expect a positive coefficient on
each of the individual Big 5 indicator variables for two reasons. First, each of the Big 5 auditors has
expanded into the consulting industry and has aggressively marketed these services to their clients
(Brynes et al. 2002). Second, Big 5 auditors have a much greater number of clients over which to
amortize the relatively fixed costs of consulting training, suggesting a greater amount of resources
devoted to the development of NAS expertise (Firth 1997). Finally, we also investigate the impact of
interactive effects (between client size and Big 5 auditor) on the NAS fee ratio. While many regula-
tors have voiced concerns about the potentially compounding effect of client size on the Big 5’s

11
Parametric and nonparametric tests indicate the difference between auditor groups (Big 5 versus non-Big 5) in terms of the
frequency of NAS fees is significant (p < .05).

Accounting and the Public Interest Volume 3, 2003

Abbott, Parker, Peters, and Rama 14

TABLE 4
Nonaudit Fee Ratios (Frequencies)

Panel A: Frequencies by Client Size

NAS Fee Ratios
Size Quartile 0.00 0.01 – 0.50 0.51 – 1.00 1.01 – 2.00 2.01 – 3.00 > 3.00
< $79.86 69 316 99 166 25 24
(n = 699) (9.9%) (45.2%) (14.2%) (23.7%) (3.6%) (3.4%)
$79.86 to 27 197 144 189 72 69
$317.95 (3.9%) (28.2%) (20.6%) (27.1%) (10.3%) (9.9%)
(n = 698)
$317.96 to 15 161 152 170 92 109
$1381 (2.1%) (23.0%) (21.7%) (24.3%) (13.2%) (15.6%)
(n = 699)
> $1381.01 8 76 102 219 108 186
(n = 699) (1.1%) (10.9%) (14.6%) (31.3%) (15.5%) (26.6%)
Total 119 750 497 744 297 388
(n = 2795) (4.3%) (26.8%) (17.8%) (26.6%) (10.6%) (13.9%)
Size quartiles = Total Assets (millions). The percentages may not add to 100 due to rounding.

Panel B: Frequencies by Auditor

NAS Fee Ratios
Firm 0.00 0.01 – 0.50 0.51 – 1.00 1.01 – 2.00 2.01 – 3.00 > 3.00
Andersen 16 93 100 124 59 79
(n = 471) (3.4%) (19.7%) (21.2%) (26.3%) (12.5%) (16.8%)
EY 10 109 109 135 68 61
(n = 492) (2.0%) (22.2%) (22.2%) (27.4%) (13.8%) (12.4%)
DT 12 98 78 92 24 44
(n = 348) (3.4%) (28.2%) (22.4%) (26.4%) (6.9%) (12.6%)
KPMG 16 105 71 77 37 47
(n = 353) (4.5%) (29.7%) (20.1%) (21.8%) (10.5%) (13.3%)
PWC 13 119 106 152 71 12
(n = 573) (2.3%) (20.8%) (18.5%) (26.5%) (12.4%) (19.5%)
Non-Big 5 41 120 44 26 6 4
(n = 241) (17.0%) (49.8%) (18.3%) (10.8%) (2.5%) (1.7%)
Total 108 644 508 606 265 347
(n = 2478) (4.4%) (26.0%) (20.5%) (24.4%) (10.7%) (14.0%)
The percentages may not add to 100 due to rounding. The Panel B sample size equals 2,478. This equates to 2,795
less 317 firms that did not have auditor data available on Compustat.

Accounting and the Public Interest Volume 3, 2003

Abbott, Parker, Peters, and Rama 15

consulting fees, we have no a priori prediction about the relation between the interactive effect and
the NAS fee ratio.
Table 5 shows a regression of the NAS fee ratio on client total assets, indicator variables for
the Big 5 accounting firms, and interactive terms between client size and each of the Big 5 auditor
indicator variables. Due to the inclusion of some very large financial institutions in our sample, all
continuous variable values greater than the 95th and less than 5th percentiles were winsorized to
equal their respective 95th and 5th percentile value of the distribution.12 Since our regressions
involve individual Big 5 auditor indicator variables, the non-Big 5 auditors represent a “comparison
set” in which to interpret the signs and magnitudes of the Big 5 coefficient estimates. Each of the
Big 5 auditors is significantly related to the magnitude of the NAS fee ratio, consistent with our
univariate results. The magnitude of coefficients is also consistent with our univariate results. In
particular, PricewaterhouseCoopers and Andersen have the largest coefficient estimates and Deloitte
& Touche the smallest.13

TABLE 5
Regression Results

NASFEERATIO = b0 + b1 Client Assets + b2 AA + b3 DT + b4EY + b5KPMG + b6PWC

+ b7Client Assets*AA + b8 Client Assets*DT + b9 Client Assets*EY
+ b10Client Assets*KPMG + b11 Client Assets*PWC + e

Expected Parameter Standard

Variable Name Sign Estimate Error p- value
Intercept 0.9324 0.0935 0.0001
Client Assets (in 000s) + 0.0022 0.0003 0.0001
Arthur Andersen (AA) + 0.9234 0.1379 0.0001
Deloitte & Touche (DT) + 0.5461 0.1511 0.0003
Ernst & Young (EY) + 0.6963 0.1355 0.0001
KPMG + 0.5929 0.1475 0.0001
PricewaterhouseCoopers (PWC) + 0.9675 0.1302 0.0001
Client Assets * AA ? –0.0074 0.0036 0.0413
Client Assets * DT ? –0.0188 0.0048 0.0001
Client Assets * EY ? –0.0097 0.0031 0.0001
Client Assets * KPMG ? –0.0211 0.0044 0.0001
Clients Assets * PWC ? –0.0121 0.0052 0.0214
Adjusted R2 = 0.112 # of obs. = 2,478
All continuous variable values greater than the 95th and less than 5th percentiles were winsorized to equal their
respective 95th and 5th percentile value of the distribution.

12
These approximately 20 financial institutions—representing less than 1 percent of the sample—introduce tremendous skew-
ness in the distribution of client assets. For example, Citibank (our largest client firm) has total assets of $902 billion, but our
95th percentile firm has $6 billion in total assets, or approximately 1/150th of Citibank’s total assets.
13
Tests of differences in regression coefficients found significant differences (p < .05) between the two NAS ratio leaders
(PricewaterhouseCoopers and Andersen) and the remaining Big 5 firms. However, intra-group comparisons (e.g.,
PricewaterhouseCoopers versus Andersen) did not yield significant differences between coefficient estimates.

Accounting and the Public Interest Volume 3, 2003

Abbott, Parker, Peters, and Rama 16

In terms of the impact of client size on the NAS fee ratio, the only class of audit firms for which
the relation is in the positive and predicted direction is for the non-Big 5 firms. This is because the
coefficient estimate on size, b 1, represents the impact of client size on the NAS fee ratio for non-Big
5 audit firms. In order to obtain the impact on client size on the NAS fee ratio for each of the Big 5
audit firms, one must add the coefficient estimate on size (b 1) with the respective interactive term
coefficients (b7–b 11). It appears that the impact of client size on the NAS fee ratio is negative for the
Big 5 firms. This is because the coefficient estimate on size is 0.0022, whereas each of the interac-
tive term coefficients b 7–b 11 is not only negative, but has an absolute value in excess of 0.0022.
Thus, there does not appear to be a compounding effect of client size on the NAS fee ratio for Big
5 audit firms vis-à-vis non-Big 5 firms. In other words, client size does not appear to escalate the
NAS fee ratio for Big 5 audit firms, after controlling for the Big 5/non-Big 5 dichotomy. This suggests
that fears regarding the impact of client size on the Big 5’s propensity to offer NAS to their clients
appear unwarranted.
In closing we note that our regression model explains only approximately 11 percent of the
variation in NAS fees, which confirms a need for further examination of this issue. Such studies
include (but are not limited to) investigating the impact of anticipated public disclosure on the NAS
fee ratio (Parkash and Venable 1993), the impact of the new Sarbanes-Oxley Act on the NAS fee
ratio, and the relation of audit committee characteristics to the NAS fee ratio.14

Comparison with Some Prior Studies

Table 6 provides data from some prior studies that examined nonaudit purchases by compa-
nies from their incumbent auditor. All of the prior studies listed in the table examined the period
from 1978 through 1981, the last time audit fees and nonaudit fees were required to be disclosed.15

TABLE 6
Comparison with Prior Studies
Mean Median Percent of companies
Study NAS Fee Ratio NAS Fee Ratio with NAS ratio > 50%
Scheiner and Kiger (1982) NA NA 11.3
Glezen and Millar (1985)a 0.30 NA NA
Beck et al. (1988)b 0.31 0.20 13.0
Parkash and Venable (1993) 0.24 0.17 9.9
Current Study: 1.60 0.96 72.6
a
Glezen and Millar (1985) provide data for three separate years (1979 through 1981) and the mean ranges from 29.2
percent to 30.7 percent for the three years.
b
Beck et al. (1988) provide data for two years (1978 and 1979), but we take the higher of the two (1978) for the
purposes of this study.

14
We also examined the impact of industry membership on the NAS fee ratio. To accomplish this, we stratified firms into 12 focus
industries per Hogan and Jeter (1997). The two industries with the largest mean NAS fee ratios were Information and Commu-
nication Services (two-digit SIC codes 48, 73, 78, 79, 84) and Real Estate Service (two-digit SIC codes 65 and 70). These ratios
were 2.171 and 2.569, respectively.
15
Each of these prior papers aggregated all nonaudit fees per ASR No. 250 when calculating their nonaudit-fee-to-audit-fee ratio.

Accounting and the Public Interest Volume 3, 2003

Abbott, Parker, Peters, and Rama 17

The last row of the table provides the results from the current study. Whether it is the mean (me-
dian) nonaudit fee ratio or the proportion of companies with NAS fee ratio higher than 50 percent,
the numbers from the current study are substantially higher than the results reported in prior stud-
ies. Parametric and nonparametric tests indicate the difference between time periods in terms of
the relative magnitude of NAS fees is significant (p < .05).16 Clearly, the nonaudit service environ-
ment has changed dramatically in the past two decades and additional research related to nonaudit
services is essential.

Financial Information Systems Design and Implementation

The SEC initially proposed prohibiting the performance of financial information systems de-
sign and implementation by auditors to their SEC audit clients. However, in the final rules, the SEC
relented and did not prohibit such nonaudit work by auditors. Instead, the final rules required com-
panies to disclose in their proxy statements the fees paid for such nonaudit work as a separate line
item.
Given the above findings with respect to the overall level of total NAS fees, we also examined
the number of companies that engaged their auditors for FIS. Fifteen percent of the sample firms
indicated that they had nonzero amounts for financial information systems design and implementa-
tion (FIS) fees. The FIS/audit fee ratio ranged from 0.012 for the smallest quartile of firms to the
0.33 for the largest quartile (Table 2, Panel B). Per Table 3, the highest mean Big 5 FIS fees were
received by PWC ($347,989), and lowest by Ernst & Young ($113,851). However, examining the
absolute magnitude of the fees indicated that $558 million of the total $3.7 billion received by
auditors for nonaudit work came from FIS fees. Fees received for FIS alone amount to 35 percent
of the total audit fees ($558 million/$1.6 billion) received from our sample of companies. Among the
15 percent of companies with FIS fees, the mean fee was $2.5 million.
These results indicate that while the frequency of financial information systems design and
implementation by auditors for their SEC audit clients was low, the magnitude of such fees was
generally quite large. This suggests that the SEC’s special disclosure emphasis related to financial
information systems design and implementation work was well warranted. However, given that FIS
engagements tend to be sporadic and project-oriented, these results are not necessarily surprising.

CURRENT DEVELOPMENTS
The SEC’s post-February 5, 2001 comments suggest that the SECPS data was the primary
source of information used by the SEC during the rule-making process. During her June 25, 2001
speech titled “This Year’s Proxy Season: Sunlight Shines on Auditor Independence and Executive

16
In cases where authors provided both mean and standard deviation information (i.e., Parkash and Venable 1993; Beck et al.
1988), we performed a parametric, two-sample t-test for differences in means. In cases where authors provided mean or
median information, without underlying standard deviation information, we performed the nonparametric, one-sample sign test
(i.e., Glezen and Millar 1985). Finally, when comparing ratio distributions to prior studies (i.e., the percentage of firms with NAS
greater than 50 percent of audit fees), in all cases we used the nonparametric, one-sample sign test. In terms of individual
auditor comparisons between reporting regimes, we can compare only to Parkash and Venable (1993), as this is the only prior
study that provides auditor-specific information concerning the NAS fee ratio, as well as standard deviation information. In all
cases, our t-tests indicate a significant increase in the NAS fee ratio between reporting regimes at p < .05. In cases where
mergers occurred in the interim between reporting regimes (i.e., Ernst & Young, Deloitte & Touche, PricewaterhouseCoopers),
we compared our current NAS fee ratio information to both predecessor firms. For example, our Ernst & Young NAS fee ratio
information was compared to Parkash and Venable’s (1993) Arthur Young and Ernst & Whinney fee ratio information.

Accounting and the Public Interest Volume 3, 2003

Abbott, Parker, Peters, and Rama 18

Compensation,” Laura Unger (interim SEC Chair before Harvey Pitt’s appointment) stated that “we
were all quite surprised to learn this … ” and “[i]n fact, the numbers appear to demonstrate the
problem may be larger than we originally thought.” These comments suggest that: (1) the SEC was
relying on the SECPS data, and (2) the SEC was surprised by both the frequency and magnitude of
the provision of NAS (Unger 2001).
The recent Enron hearings have furthered the public scrutiny surrounding auditor indepen-
dence. Many have called for the outright banning of all consulting work performed by incumbent
auditors (Brynes et al. 2002). Still others have called for greater audit committee involvement in the
NAS purchase decision.17 The events surrounding Enron’s failure have caused the Big 5 to re-
assess their stance toward the provision of NAS to their audit clients. Recently, KPMG and PWC
have announced plans to spin off their consulting divisions. This is in addition to Ernst & Young,
which sold its consulting division to Cap Gemini during 2001. Interestingly, Arthur Andersen, the
one Big 5 firm that had spun off its consulting division in August 2000, had the second highest (third
highest) mean (median) NAS fee ratio. However, the proxy statement disclosures themselves did
not specify whether or how much of the nonaudit fees paid to Andersen in 2000 related to their
newly spun-off Accenture consulting firm. Consequently, it remains an empirical question how much
impact these spin-offs will have on the NAS fee ratio.
The events of Enron were compounded by the recent WorldCom scandal. The combination of
events gave rise to the recently enacted Sarbanes-Oxley Act. The Sarbanes-Oxley Act, effective in
2003, bans certain NAS, but excludes tax services from the list of proscribed services. The banned
NAS include: internal audit outsourcing, bookkeeping, or other services related to the accounting
records or financial statements of the audit client; financial information systems design and imple-
mentation; appraisal or valuation services; actuarial services, management functions, or human
resources; broker or dealer, investment adviser, or investment banking services; and legal and
expert services unrelated to the audit. Interestingly, the newly banned NAS are ostensibly identical
to the ones the SEC originally sought to proscribe while facing severe criticism by prior legislative
bodies and a large sector of public accounting.
The Sarbanes-Oxley Act also created a new accounting governing body entitled the Public
Company Accounting Oversight Board. The new oversight body will not be funded by the AICPA
nor Big 5 accounting firms, rather it will be funded by the SEC registrants themselves. The Big 5
and other accountants that audit public companies will be required to comply with the board’s
requests for evidence and will be subject to its discipline. The Board will have subpoena power and
is scheduled to have a majority of its members to be nonpractitioners.

SUMMARY AND CONCLUSIONS

During the recent rule-making process related to nonaudit services, the SEC relied on statis-
tics provided by the firms to the SECPS in describing the magnitude and frequency of NAS pur-
chases. However, the SECPS data are incomplete for the NAS debate because they provide infor-
mation only about a subset of NAS, namely management consulting services. Results from our
analysis of proxy statements filed by companies under the new SEC rules indicate that only about
4 percent of companies did not purchase NAS from the incumbent auditor, as opposed to the 75

17
This is a particularly interesting topic given that both the SEC and the AICPA agree the audit committee should play a more
active role in the NAS purchase decision.

Accounting and the Public Interest Volume 3, 2003

Abbott, Parker, Peters, and Rama 19

percent estimated by the SEC. Further, 51 percent of the companies paid NAS that were greater
than the audit fee, as opposed to the “less than 5 percent” estimated by the SEC.
Together, the above findings suggest that the SEC may have significantly underestimated both
the magnitude and frequency of NAS. Interim SEC Commissioner Laura Unger (2001) noted the
disparity between actual NAS purchases and that from the SECPS was “eye-opening, to say the
least.” Lynn Turner (2001), former SEC chief accountant, has noted “it’s a whole new ball game.” At
a minimum, our results suggest the public interest impact of NAS fees remains an unresolved issue
that demands further research.
Aside from our call for future research, there are at least two public policy implications of our
paper. First, the SEC should maintain, at the very minimum, the current disclosure environment. It
will allow researchers the opportunity to empirically study some of the issues raised in this paper.
Preliminary studies have, thus far, yielded mixed results in this regard. Frankel et al. (2002) find that
the NAS-audit fee ratio is positively associated with the abnormal accruals and the ability to meet
analyst forecasts, whereas DeFond et al. (2002) find a positive relation between the NAS-audit fee
ratio and the propensity to issue a going-concern opinion. Second, a more informative disclosure
would likely assist marketplace participants to better assess the overall client-auditor relationship.
It would be worthwhile to determine the investing public’s preference for disclosure format regard-
ing FIS and other NAS fees. Furthermore, the definitions of “audit-related” nonaudit fees should be
standardized and a more detailed description of the NAS engagements should be disclosed.
In sum, the size and magnitude of NAS purchases suggests a public interest issue of great
concern to investors, accountants, regulators, and academics. The speed in which the Sarbanes-
Oxley Act was enacted suggests a dramatic and swift change in the perception of NAS on auditor
independence. Nonetheless, there remains a host of empirical questions that must be addressed
before any conclusions can be drawn. However, academics are in a uniquely neutral position to
study issues related to auditor independence and nonaudit services. We encourage further re-
search in this area as it serves the public interest to know more about the issues raised in this
paper.

REFERENCES
Arthur Young & Company. 1978. Disclosing Relations with Independent Auditors—ASR 250: Toward Understanding
and Compliance. New York, NY: Arthur Young & Co.
Barkess, L., and R. Simnett. 1994. The provision of other services by auditors: Independence and pricing issues.
Accounting and Business Research 24 (94): 99–108.
Beck, P. J., T. J. Frecka, and I. Solomon. 1988. An empirical analysis of the relationship between MAS involvement and
auditor tenure: Implications for auditor independence. Journal of Accounting Literature 7: 65–84.
Brynes, N., M. McNamee, D. Brady, and L. Lavelle. 2002. Accounting in crisis. BusinessWeek (January 28): 44–48.
Carney, D., and L. Cohn. The swamp fox on Enron’s tail. BusinessWeek Magazine. (February 4): 36–37.
DeFond, M., K. Raghunandan, and K. R. Subramanyam. 2002. Do nonaudit services impair auditor independence?
Evidence from going-concern audit opinions. Working paper, University of Southern California, Los Angeles, CA.
Ernst & Young. 2001. Financial Reporting Briefs. Chicago, IL: Ernst & Young, LLP.
Firth, M. 1997. The provision of nonaudit services by accounting firms to their audit clients. Contemporary Accounting
Research (Summer): 1–21.
Frankel, R., M. Johnson, and K. Nelson. 2002. The relationship between auditor’s fees for non-audit services and
earnings management. The Accounting Review 77 (Supplement): 71–105.
Glezen, G. W., and J. A. Millar. 1985. An empirical investigation of stockholder reaction to disclosures required by ASR
No. 250. Journal of Accounting Research (Autumn): 859–870.
Hogan, C., and D. Jeter. 1997. Industry specialization by auditors. Working paper, Vanderbilt University, Nashville, TN.

Accounting and the Public Interest Volume 3, 2003

Abbott, Parker, Peters, and Rama 20

Levitt, A. 1996a. The guardians of financial truth. Remarks delivered at the Financial Reporting Institute, University of
Southern California, Pasadena, CA, June 6.
———. 1996b. The accountant’s critical eye. Remarks delivered at the 24th Annual National Conference on Current
SEC Developments, American Institute of Certified Public Accountants, Washington, D.C.
———. 1998. The numbers game. Remarks delivered at the NYU Center for Law and Business, New York, NY,
September 28.
———. 1999. Remarks delivered to the Panel on Audit Effectiveness of the Public Oversight Board, Public Oversight
Hearings, New York, NY, October 7.
———. 2000. Renewing the covenant with investors. Speech before the NYU Center for Law and Business, New York,
NY, May 10.
Liesman, S., J. Weil, and M. Schroeder. 2002. Dirty books? Accounting debacles sparks calls for change: Here’s the
rundown. Wall Street Journal (February 6): A1.
Lowe, D. J., M. Geiger, and K. Pany. 1999. The effects of internal audit outsourcing on perceived external auditor
independence. Auditing: A Journal of Practice & Theory 18 (Supplement): 7–26.
McCloy, J. J. 1979. Letter to Walter E. Hanson, Chairman, Executive Committee, SEC Practice Section. March 9.
McNamee, M. 2002. Pitt’s accounting fix leaves a lot broke. BusinessWeek (February 24).
Messier, W. F. 2000. Auditing and Assurance Services: A Systematic Approach. New York, NY: McGraw-Hill.
Palmrose, Z-V. 1986. The effect of nonaudit services on the pricing of audit services: Further evidence. Journal of
Accounting Research 24 (Autumn): 405–421.
Pany, K., and P. M. J. Reckers. 1983. Auditor independence and nonaudit services: Director views and their policy
implications. Journal of Accounting and Public Policy (Spring): 43–62.
———, and ———. 1984. Nonaudit services and auditor independence: A continuing concern. Auditing: A Journal of
Practice & Theory (Spring): 89–97.
Parkash, M., and C. Venable. 1993. Auditee incentives for auditor independence: The case of nonaudit services. The
Accounting Review (January): 113–134.
Roberts, R. W. 2001. Commercialism and its impact on the integrity of professional tax services in the United States.
Critical Perspectives in Accounting 12: 589–605.
Scheiner, J. H., and J. E. Kiger. 1982. An empirical investigation of auditor involvement in nonaudit services. Journal of
Accounting Research (Autumn): 482–496.
Securities and Exchange Commission (SEC). 1982. SEC Accounting Rules. Chicago, IL: Commerce Clearing House.
———. 2000a. Proposed rule: Revision of the commission’s auditor independence requirements. 17 CFR Parts 210
and 240. Release Nos. 33-7870;34-42994; 35-27193; IC-24549; IA-1884, File No. S7-13-00. Washington, D.C.:
Government Printing Office.
———. 2000b. Revision of the commission’s auditor independence requirements. Release Nos. 33-7919; 34-43602;
File No. S7-13-00. Washington, D.C.: Government Printing Office.
———. 2000c. Hearing on auditor independence. September 20. Available at: http://www.sec.gov/rules/extra/
audmin3.htm.
———. 2000d. Hearing on auditor independence. September 21. Available at: http://www.sec.gov/rules/extra/
audmin4.htm.
Shad, J. 1981. Comments appearing in “Current News Section.” Accountancy Magazine (October): 9–10.
Swanger, S., and E. Chewning. 2001. The effect of internal audit outsourcing on financial analysts’ perceptions of
external auditor independence. Auditing: A Journal of Practice & Theory 20 (September): 115–129.
Turner, L. 2001. Statements made at the American Accounting Association Annual Meeting, Philadelphia, PA, August
21.
Unger, L. 2001. This year’s proxy season: Sunlight shines on auditor independence and executive compensation. June
25. Available at: http://www.sec.gov/news/speech/spch502.htm.

Accounting and the Public Interest Volume 3, 2003