Separation of tax service from audit service, going‐concern opinions, and discretionary accruals

• Published date: 01 April 2022
• Author: Kam‐Wah Lai
• Date: 01 April 2022
• DOI: http://doi.org/10.1002/jcaf.22532

Received: 28 September 2021Accepted: 8 December 2021

DOI: 10.1002/jcaf.22532

RESEARCH ARTICLE

Separation of tax service from audit service, going-concern

opinions, and discretionary accruals

Kam-Wah Lai

Department of Accounting and Banking,

Chu Hai College of Higher Education,

Tuen Mun, Hong Kong

Correspondence

Kam-WahLai, Department of Accounting

andBanking, Chu Hai College of Higher

Education,Tuen Mun, Hong Kong. Email:

kwlai@chuhai.edu.hk

Abstract

This paper examines whether separating tax service of auditors from audit

service improves audit quality. To some critics, the separation is the only

solution to the independence problem created by the joint provision of both

services by auditors. However, opponents argue that auditors are likely to have

less client-specific knowledge when they only provide audit service, resulting

in lower audit quality. The results of this paper show that clients without the

purchase of auditor-provided tax service are more likely to havea going-concern

modified audit opinion and lower discretionary accruals than clients with such

purchase. Additional tests indicate that clients with (without) auditor-provided

tax service would have had lower (higher) discretionary accruals or been more

(less) likely to receive a going-concern opinion had they had not (had) purchased

the tax service. The results are robust to the control for possible self-selection

problem of clients’ purchase of tax service of auditors.

KEYWORDS

auditor-provided tax service, discretionary accruals, going-concern opinions, joint provision

1 INTRODUCTION

During the 1990s, non-audit services became a much more

important source of revenue to audit firms1and regula-

tors were then concerned about the effect of the provi-

sion of non-audit services on audit quality because audit

firms are seen to depend tremendously on non-audit rev-

enue with possible impairment of auditor independence.

Arthur Levitt, the then chairman of the Securities and

Exchange Commission (SEC), commented that “(a)s audit-

ing becomes an ever-smaller portion of a firm’s business

with an audit client, it becomes harder to assume that

the auditor will challenge management when he or she

should, if to do so might jeopardize a lucrative consulting

contract for the auditor’s firm” (Levitt, 2000). Therefore,

academic studies use the level of non-audit fees (or non-

audit fees in a fee ratio) to study independence or audit

quality issues.

Following the Enron debacle, the Sarbanes-Oxley Act of

2002 (SOX) banned the auditors’ provision of major types

of non-audit services that were previously provided for

audit clients. This ban represents a shift away from the

level of non-audit fees because it ignores the materiality

of the amount spent. However, academic studies continue

to focus on the level of fees to examine independence or

audit quality issues (e.g., De Simone et al., 2015; Seethara-

man et al., 2011). In fact, regulators later de-emphasize the

level of fees and focus more on the provision of non-audit

services itself. The Public Company Accounting Oversight

Board (PCAOB) suggested that “(e)ven when an audit firm

does not provide significant non-audit services to a partic-

ular issuer audit client, we are concerned about the effects

such business developments may have on the firm’s atten-

tion to audit quality” (Public Company Accounting Over-

sight Board (PCAOB), 2015). In the United Kingdom, reg-

ulators proposed to require audit firms to perform audits

J Corp Account Finance. 2022;33:13–30. © 2021 Wiley PeriodicalsLLC13wileyonlinelibrary.com/journal/jcaf

14 LAI

only for clients, following the collapse of Carillion, a con-

tractor (White, 2020) because “.. . separating accountancy

firms’ audit departments from the rest of their operations

would protect auditors ‘from influences from the restof the

firm that could divert their focus away from audit quality’”

(BBC News, 2020, comments of the Financial Reporting

Council in quotation marks in original).

Therefore, there is a need for academic studies toaddress

regulatory concerns by studying the separation of audit

and non-audit services of auditors but not just the level of

fees. The evidence (which is mixed) that audit quality is

not reduced by high non-audit fees does not address the

request for separating the services by critics of joint provi-

sion.2This issue of separation is interesting because “(t)he

delivery of a high quality audit requires audit teams who

are able to access a broad base of specialist knowledge .. .

This could be difficult to maintain in a firm which focussed

only on audit services” (Pickard, 2018). Thus, whether the

separation will improve audit quality remains an empirical

question.

This paper investigates the effect of the separation of tax

and audit services on audit quality. It tests whether audit

clients without auditor-provided tax service are more, or

less, likely to have a going-concern modified audit opin-

ion or higher, or lower, discretionary accruals than clients

with such tax service.3The use of the two measures of audit

quality is consistent with and hence, addresses the con-

cern of regulators that “.. .paying tax fees to audit firms .. .

could compel auditors to acquiesce to clients’ aggressive

financial reporting and issue unqualified audit opinions”

and such payment “.. . could result in lenient auditing of

any account, not just the tax accounts” (Cook et al., 2020,

p. 84, emphasis in italics in original). Furthermore, tax ser-

vice of auditors deserves attention in its own right because

it created independence concern long before and shortly

after SOX4and the “.. . economic-bonding argument sug-

gests that revenue generated by the provision of any non-

audit service to the client could result in more lenient

auditing ...” (Cooket al.,2020, p. 84, emphasis in italics

in original).

Using U.S. data, this paper reports that clients with-

out auditor-provided tax service are more likely to have

a going-concern opinion and lower discretionary accru-

als than clients with such service, with economic effect of

12.57% and 6.04%, respectively for an average firm. Addi-

tional tests show that if clients with tax service of auditors

had not had purchased the service, they would have been

more likely to receive a going-concern opinion and have

lower discretionary accruals. The converseholds for clients

without auditor-provided tax service. Last, the results are

robust to the control for the possible self-selection issue of

the purchase of auditor-provided tax service. Overall, the

evidence suggests that audit quality is higher with the sep-

aration of audit and tax services of auditors.

There are some studies on the separation of tax service

of auditors but except for Cripe and McAllister (2009), they

focus on the benefits of the service to the clients (i.e., the

tax outcome) but not on the effect on audit service.5Using

t-tests for firms from 2003–2005, Cripe and McAllister

(2009) do not find significant differences in discretionary

accruals between clients with and clients without auditor-

provided tax service. In contrast, this paper, using a larger

sample from 2004–2017 and regression analyses, provides

new and more recent results that clients without tax ser-

vice of auditors have lower discretionary accruals than

clients with the service. Another difference is that Cripe

and McAllister (2009) do not study the issuance of a mod-

ified audit opinion but this paper shows that clients with-

out auditor-provided tax service is more likely to receive a

going-concern opinion.

This paper makes the following contributions. Its results

show that the separation of audit and tax services of audi-

tors improves audit quality. Since tax service is a major

non-audit service allowable by regulators (e.g., the SEC),

these results provide good inference to support the sugges-

tion that the separation of non-audit and audit services will

enhance auditor independence. As argued by some crit-

ics (e.g., Mautz & Sharaf, 1961), without the joint provi-

sion of audit and non-audit services, audit firms will not

depend on non-audit revenue and are not under the atten-

dant threats to independence. Therefore, this paper con-

tributes to the debate of adopting the audit-only model for

audit firms and deserves the attention of regulators, audit

clients, auditors and users of financial statements.

Second, this paper adds to the literature on audit qual-

ity by providing evidence in a different setting. While prior

studies use the level of non-audit fees to study auditor

independence or audit quality issues, this paper addresses

whether the provision of auditor-provided non-audit ser-

vices itself reduces audit quality.6Such an issue becomes

increasingly more pertinent given regulatory initiatives

to improve auditor independence following the continual

occurrence of accounting scandals worldwide.7

Section 2of this paper develops the hypothesis. Section 3

explains the research method and the sample selection.

Section 4discusses the results, including those of addi-

tional tests and the last section concludes the findings.

2 HYPOTHESIS

SOX allows the provision of tax service by incumbent audi-

tors because “the review of the registrant’s tax reserves

requires substantial knowledge about the audit client”