Consequences of the Abandonment of Mandatory Joint Audit: An Empirical Study of Audit Costs and Audit Quality Effects

Abstract

This paper focuses on the unique Danish setting in examining the consequences of abandoning a mandatory joint audit regime. We study the effects on audit costs (measured by audit fees) and audit quality (measured by abnormal accruals) of the abandonment of the mandatory joint audit in Denmark in 2005. We perform our analysis on non-financial listed Danish companies for the 2002–2010 period. Our results show that a joint audit is associated with higher fees, but that the association between joint audit and abnormal accruals is insignificant. This suggests that the higher audit fees cannot be explained by higher audit quality. Our results are robust to alternative measurements of fees and audit quality. Additional analyses show that the fee premium related to a joint audit decreases over time and that the Big 4 concentration in our sample has increased since the switch from mandatory to voluntary joint audit. Our results are consistent with the motivations driving the regulatory change in Denmark and are of interest to regulators and actors in the audit market.

Acknowledgements

The authors thank Walid Alissa, Denise Hanes, Clive Lennox, Donald Stokes as well as the participants of a workshop held at the conference Audit Market Structure, Competition and Audit Quality (KU Leuven, Sept. 2011) for comments on earlier versions of this paper. We also gratefully acknowledge the comments provided by participants of the following conferences and workshops: the EAA conference (May 2012); workshops at Bentley University (March 2012), Concordia University (April 2012) and Humboldt University (November 2012); the 4th Workshop on Audit Quality (September 2012); and the Midyear Conference of the American Accounting Association (January 2013). In addition, we thank Emna Ben Saad for her valuable assistance. Finally, we thank the editor, Ann Vanstraelen, and one anonymous reviewer for their extremely valuable comments and insights.

Disclosure Statement

No potential conflict of interest was reported by the authors.

Notes

¹The Big 4 market share 2010 for companies with FTSE 350 equivalent market cap was 58%.

²An agreement between the European Parliament and the EU Member States on the reform of the audit market was finally reached (European Commission, 2013) whereby joint audit is encouraged. In the final EU regulation about the specific requirements regarding the statutory audit of public-interest entities, the period for mandatory audit firm rotation is at maximum ten years, with the member state option to shorten this period; after a maximum of ten years, the period can be extended by an additional ten years if tenders are carried out, and for an additional 14 years in the case of joint audits. This extension of the duration of the audit mandate is also a member state option (European Union, 2014, Art. 17).

³In their defense, the authors do acknowledge this limitation and seek to address it.

⁴However, the EU study of systems of civil liability of statutory auditors identified that it is possible in Denmark, France, Ireland and Sweden to apportion the liability according to the actual individual involvement under specific conditions (EC, 2001, p. 32). In Denmark, an audit firm that is not negligent in a joint audit may avoid liability (see for further details Holm & Thinggaard, 2015).

⁵In addition, we also detected inconsistencies in the audit/non-audit fee categorization across companies. In Denmark, the companies disclosed fees across two categories (audit, non-audit) in the years 2002–2007, while disclosing them across four categories (audit, other assurance services, tax advisory, other advisory) from 2008 onwards. We found examples of inconsistent allocation across companies of the ‘other assurance services' between audit and non-audit categories. For example, Andersen and Martin categorized ‘other assurance services' under non-audit fees, while ALK Abello considered them to be audit fees (source: companies' financial statements from 2008 to 2009). Thus, using both audit fees and total fees appears to be an adequate approach to capture the various concerns of the fee categorization across companies.

⁶We also use another audit quality proxy, namely earnings benchmark tests, in an additional analysis (see Section 7).

⁷We also use an alternative model to calculate the abnormal accruals in an additional analysis (see Section 7).

⁸In its response to the EC Green Paper, Grant Thornton mentioned that 16 out of the 64 largest Danish public companies still preferred joint audit. This proportion has often been cited in the EC debate on joint audit. Despite our inquiries made to the EC and professional organizations, we have not been able to get the list of these companies. Our own investigation found that: (1) this proportion may refer to a pre-2009 situation (e.g., 10% of companies were jointly audited in our 2008 sample); (2) most of these companies may be banks or insurance companies, or unlisted state-owned companies; and (3) following our request by email, a representative of the professional body of the Danish auditors has confirmed that none of the 100 Danish largest listed companies are still jointly audited in 2015.

⁹Note that audit market concentration is determined by diverse mechanisms and that the role of joint audit (or its abandonment) is extremely difficult to disentangle in empirical research. For example, the client base of the Big 4 audit firms in Denmark continued to increase in 2011 when Grant Thornton Denmark was acquired by PwC.

¹⁰The abnormal accruals measures are winsorized at 0.01 after the calculation of the corresponding measures.

¹¹Due to this high correlation between our test variable JOINT and the control variable IFRS we have split the sample in companies that have adopted IFRS and those that have not adopted IFRS in an additional test (see Section 7).

¹²We are not able to include an industry effect because of an insufficient number of data per industry segment in Denmark. However, when we include dummy variables for the most frequent industries in our sample – Manufacturing sectors (SIC 20-39) –, results (untabulated) remain unchanged.

¹³The percentage is given by the following transformation: %=exp (β₁)-1.

¹⁴The Danish setting does not allow us to use other proxies, such as going concern opinions or restatements, due to the limited occurrence of those events and the resulting lack of sufficient variance for these variables.

¹⁵The control variable vector is the same as in model (2), with the exception of the LOSS_LAG, which is dropped.

¹⁶See for instance the rejection by Salustro (a French second-tier audit firm) of a fraudulent accounting treatment made by Vivendi in 2002 and allowed by Arthur Andersen. The AMF (the French stock market regulator) eventually supported Salustro's position (Davis, Lukomnik, & Pitt-Watson, 2006, p. 135).

Additional information

Funding

Cédric Lesage acknowledges the financial support of the HEC Foundation (Project F1103) and the INTACCT program (European Union, Contract no. MRTN-CT-2006-035850). He is a member of GREGHEC, CNRS Unit, UMR 2959. Jaana Kettunen acknowledges the financial support of the Finnish Foundation for Economic Education.