Paying for Joint or Single Audits? The Importance of Auditor Pairings and Differences in Technology Efficiency

• Date: 01 March 2016
• Author: Frank Thinggaard,Claus Holm
• DOI: http://doi.org/10.1111/ijau.12050
• Published date: 01 March 2016

Paying for Joint or Single Audits? The Importance of Auditor Pairings and

Differences in Technology Efficiency

Claus Holm and Frank Thinggaard

Aarhus University

In the first theoretical paper on joint audits,Deng et al. predict that the audit fees for joint audits will be lower than

those from single audits. However, the prediction depends on the combination of audit firms involved in the joint

audit and on their technology efficiency as well as on the liability involved. This paperis the first to empirically test

the predictions. Our findings from Denmark do not indicate any general difference in audit fees when two audit

firms – regardless of combination and technology efficiency – conduct the statutory audit compared to a single Big

audit firm. Theresults indicate the existence of fixed coordination costs in joint audits.We do, however, find higher

audit fees in Big-Small joint audits when the Small audit firm has a share of less than 25 per cent. This may reflect

free-riding concerns.

Key words: Joint audits, auditor pairings, audit technology efficiency, audit fees, B4

1. INTRODUCTION

Many opponents of joint audits claim that single audits

are always cheaper but, according to a recent theoretical

paper by Deng et al. (2014), this is not the case. In fact,

their analysis reverses the question and identifies the

cases where audit fees for joint audits are lower than

audit fees for companies audited by a single auditor. Our

study is the first to empirically test the theoretical

predictions.

In a joint audit two audit firms work together to

produce a single audit report, thereby sharing the

responsibility for the audit. While joint audits have

generally been allowed for corporations within the EU,

they have been limited to certain special circumstances

in most member countries (EC, 2001, p. 34). In a green

paper, the European Commission recently suggested

more widespread use of mandatory joint audits and thus

responded to concerns about the overwhelming

dominance of the Big auditfirms in relation to the audit of

large companies. The Commission proposed that the

limited practice of appointing two different audit firms

should be ‘developed further to “dynamise” the market

to allow mid-tier non-systemic firms to become active

players in the market segment of the audits of large

corporations’ (EC, 2010, p. 15). The Commission’s joint

audit proposal therefore implied that a Big audit firm

should be paired with a non-Big audit firm in audits of

large corporations. In responses to the EC proposal,

opponents argued that joint audits may lead to increased

bureaucracy and added costs (EC, 2011b), which then led

to a range of surprisingly soft proposals from the

European Commission (EC, 2011a). Finally, on 3 April

2014, the European Parliament adopted an amended

directive on the statutory auditof public-interest entities.

The directive (EC, 2014) does not mandate joint audits,

but suggests the use of joint audits as a complementary

mechanism for mandatory audit firm rotation and

suggests that the maximum duration of the engagement

could be stretched from 10 to 24 years if the audit is done

jointly.

The positive attitude towards joint audits by the EU

and others (e.g., supportive joint audit responses in EC,

2011b) and the observation that single auditsare the norm

in most countries, motivated Deng et al. (2014) to

theoretically study joint audits and compare audit

evidence precision and audit fees under joint and single

audits. But Deng et al. not only analyse the two groups

(single and joint); they also identify free-riding by one of

the auditors in a joint audit as a discouraging economic

factor. They consider three regimes that may vary in the

magnitudes of potential free-riding problems (Deng et al.,

2014, p. 1030): single audits by one Big firm (Regime B);

joint audits by two Big firms (Regime BB); joint audits by

one Big firm and one Small firm (Regime BS). Deng et al.

(2014) assume that differences in technology efficiency

between a Big and a Small firm are a decisive factor when

audit fees from this (BS) joint audit pairing are compared

with a single B firm. Another assumption is that a Big

audit firm bears a larger proportion of misstatement costs

than a Small firm. Under these assumptions, Deng et al.

(2014) find that audit fees under joint audits are lower

than those under single audits if the technological

difference between the two audit firms is Small and/or

the Big firm bears a sufficiently large proportion of the

misstatement costs.

To the best of our knowledge, no prior study has

empirically tested these theoretical predictions. One

reason is that it is hard to find datathat include both joint

and single audits. Twocountries stand out with regard to

previous experience with joint audit systems. In both

France and – until recently – Denmark, joint audits have

been mandatory for certain types of statutory audits. In

France joint audits of public companies have been

mandated by law since 1966 (Ratzinger-Sakel et al., 2013).

But the absence of single audit cases makes France an

unsuitable subject for an empirical test of single versus

joint audits. Mandatory joint audits were required for

listed companies in Denmark from 1930 to 2004, but the

provision was abolished for the auditing of financial

reports after 1 January 2005 (Holm & Thinggaard, 2014).

The Danish change in 2005 from a mandatory to a

voluntary joint system led to statutory audits conducted

as single audits or joint audits.This enables us to compare

audit fees between the different auditregimes. The aim of

the paper is to empiricallytest the fee predictions in Deng

Correspondence to: Claus Holm, Department of Economics and

Business, Aarhus University, Fuglesangs Alle 4, DK-8210 Aarhus V,

Denmark. Email: hoc@asb.dk

International Journal of Auditing doi:10.1111/ijau.12050

Int. J. Audit.

© 2015 John Wiley & Sons Ltd ISSN 1090-6738

20: 1 16 (2016)

–

et al. (2014) given the Danish context. Hence, we engage

in an examination of the assumptions in Deng et al.

(2014) and the implications for empirically testing the

proposition in the setting of the Danish auditmarket. This

has led us to further examine the possible classification of

technology efficiency and cost drivers.

Our analyses are based on the core audit fee

determinants model suggested by Simunic (1980) and

used in most previous research (Cobbin, 2002; Hay,

Knechel & Wong, 2006; Hay, 2013) complemented with

auditor pairing variables divided into various technology

efficiency classifications. Our study provides three main

contributions. First, our empirical test of Deng et al.

contradicts their predictions and shows that neither

BB nor BS joint audits exhibit lower audit fees

when compared to B single audits. This implies that

coordination costs cannot be ignored and have a

significant effect on audit costs in the context of joint

audits. Second, our results show that, overall, joint audits

with at least one Big are not more expensive than single

audits performed by one Big alone. The practical

implication is that the BS joint auditregime suggested by

the European Commission can be introduced without

any negative effects on audit costs. Third, we show that,

in the case of very unbalanced BS joint audits, when the

Big auditor works with a very Small auditor taking less

than 25 per cent of the fees, the audit fees are higher than

when the Big auditor works alone on the audit. Our

interpretation of the results is that the Big audit firm is

particularly concerned about potential free-riding by the

Small audit firm when the Small audit firm has a small

share of the audit. The obvious practical implication is

that companies should take the lead and demand a more

equally shared audit if they prefer a BS joint audit.

Otherwise they should consider dismissing the Small

audit firm.

The remainder of the paper is organised as follows. In

the second section, we provide a literature review and

present the theoretical basis for our hypotheses. In the

third section, we consider technology efficiency and

liability of audit firms in the Danish setting. In the fourth

section, we describe the data, the research design and

provide descriptive statistics. We present our results on

audit fees in the fifth section. In the final section we

conclude, consider the limitations of the study and

provide proposals for future research.

2. LITERATURE REVIEW AND

HYPOTHESIS DEVELOPMENT

The analysis in Deng et al. (2014) suggests that the joint

audit combinationas well as the differences in technology

efficiency of the joint auditors must be considered when

comparing joint auditfees to single audit fees. We refer to

three different audit regimes in order to match the

exposition in their paper, that is, single audits by one Big

firm (Regime B), joint audits by two Big firms (Regime

BB) and joint audits by one Big firm and one Small/

mid-tier firm (Regime BS). Deng et al. (2014) propose that

audit fees are lower in joint audits than in single audits,

but for BS joint audits their proposition is conditioned

by the differences in technological efficiency of the

audit firms and on the Big firm’s proportion of the

misstatement costs.

Note that the potential for lower audit fees in a joint

audit setting is inconsistent with prior claims made by

audit regulators. Usually, the argument is that joint audits

will lead to higher costs due to overlapping audit parts

(EC, 2011b) and that some of the non-productive

overheads of planning, supervision and review will be

doubled when two auditors are involved. It is also

inconsistent with the cost argument provided by the

Danish legislators, who claimed that the joint audit

requirement in Denmark represented an unnecessary

financial burden on companies (Danish Financial

Statements Act, 2001, Basis for Conclusions §135).

Empirical evidence on audit fee differences between

single and joint audits is scarce (Zerni et al., 2012; Holm &

Thinggaard,2014; André et al., 2015). Similar to this study,

Holm and Thinggaard (2014) analyse audit fees in the first

years after the abolition of the Danish mandatory joint

audit system. They examine the full population of

non-financial listed companies and apply both a core

audit fee determinants model and an audit fee change

model. Holm and Thinggaard (2014) find support for fee

reductions in companies switching to single audits, but

only in the short term. The authors suggest that fee

discounts are related to competition rather than a

reflection of long-lasting efficiency gains. In additional

analyses they include interaction terms and investigate

whether the single auditor fee discount depends on how

the joint audit work was shared between the joint

auditors before the abolition. Based on these analyses the

authors find that fee reductions in companies switching

to single audits are exclusively found in situations where

the former joint audit included a dominant auditor. The

authors argue that the auditors have more bargaining

power in this situation than in case of an equally shared

joint audit and that their incentive for offering a large

initial fee discount is higher. While their study compares

audit fees in joint and single audits within a country, it

neither considers the joint audit combination, deemed

important by Deng et al. (2014), nor does their study

isolate single B audit firms like in the analysis by Deng

et al.

André et al. (2015) obtain a sample of 631 companies,

comprising 210 French companies, 279 British companies

and 142 Italian companies from 2007–11, and compare

audit fees paidby listed companies in France, where joint

audits are mandatory, with those paid by British and

Italian companies. First, they use a basic audit fee model

where they control for a number of engagement

attributes, including audit quality, and include dummy

variables which capture the potential difference in audit

fees between a French company and either a British or an

Italian one. Next, they use a more detailed model where

they distinguish between the number of Big audit firms

involved in the audit. The results indicate that joint BS

audits and joint BB audits in France are more costly than

single B audits in both the UK and Italy. The results also

indicate that French companieswith two non-B4 auditors

pay significantly more than British and Italian companies

with single auditors. The BB vs B results contradict the

predictions in Deng et al. (2014), and the BS vs B results

potentially contradict the predictions depending on the

technology efficiency status of the audit firms involved

and the Big audit firm’s proportion of misstatement risk

(see hypothesis 2 below). However, since the analysis is

based on cross-country observations, there is a risk that

different technology efficiency classifications and other

country-specific factors mayhave a confounding effect on

the results.

In Sweden, the use of single vs voluntary joint auditsin

listed companies has been examined by Zerni et al. (2012).

C. Holm and F. Thinggaard

Int. J. Audit.© 2015 John Wiley & Sons Ltd 20: 1 16 (2016)

–

2