Audit Partner Public-Client Specialisation and Client Abnormal Accruals

Abstract

We examine the association of Big 4 audit partners’ public-client specialisation with client companies’ audit quality. Using a sample of NASDAQ OMX companies in Finland, we identify the audit partner assigned to each public-client engagement. We expect that partners with greater public-client specialisation provide higher quality auditing, since they have likely developed deep domain-specific knowledge and a keen sense of the litigation and reputational risks posed by public clients. In addition, the willingness to resist client pressure likely increases with the number of public clients in the partner's portfolio because dependence on any one client diminishes, which should help to ensure audit quality. The results show that public-client specialisation is negatively associated with abnormal accruals, and this result is attributable to audit partners with three to six public clients. The results of supplemental tests imply that public-client specialisation is more important when general auditing experience is lower. Further, the results reveal that in our setting of high-tax and high alignment between financial reporting and tax reporting, greater public-client specialisation is particularly associated with smaller income-decreasing abnormal accruals, suggesting that auditors with greater public-client specialisation likely recognise the downside reputational implications and achieve audit quality by discouraging tax avoidance.

Acknowledgements

Johnstone acknowledges support from the Andersen Center for Financial Reporting at the University of Wisconsin School of Business, and the EY Foundation for support of her EY Professorship. Ittonen acknowledges financial support from the NASDAQ OMX Nordic Foundation, the Marcus Wallenberg Foundation, Southern Ostrobothnia Fund, and the Foundation for Promoting Equity Markets in Finland. Myllymäki gratefully acknowledges the financial support received from the Finnish Foundation for Economic Education, Jenny and Antti Wihuri Foundation, Nordea Bank Foundation, Oskar Öflund's Foundation, and Evald and Hilda Nissi Foundation. We appreciate the thoughtful comments of two anonymous reviewers, the editor Ann Vanstraelen, Jean C. Bedard, Marsha Keune, Nicole Ratzinger-Sakel, Stefan Sundgren, Sami Vähämaa, Mikko Zerni, and the seminar participants at the 3rd Workshop on Audit Quality (Bellagio), NHH (Bergen), and the Aarhus University Business School.

Notes

1 Abnormal accruals reflect the quality, or lack thereof, of reported earnings. Reported earnings are the output of the financial reporting and the audit process, and thus represent a signal of actual audit quality since the audit process affects the client's audited financial statements (Becker et al., 1998; Francis, 2011).

2 Audit partner signatures have been available in EU member states since the implementation of the Directive on Statutory Audit (Council of the European Union, 2006 (2006/43/EC)). However, Finland has a long tradition of publishing audit partner signatures in the annual reports.

3 We recognise that experience, knowledge, and specialisation are distinct, but related concepts, both conceptually and empirically. Experience involves ‘the learning of action-outcome connections … learning occurs through outcome feedback’ (Einhorn & Hogarth, 1981, p. 78). ‘Of great importance to the issue of learning from experience is the role of awareness of the task factors that can influence outcomes’ (Einhorn & Hogarth, 1981, p. 79.) Knowledge ‘is generally rich in content and complex in form. It includes appearance, function, relation to other objects, and any other property of the object that can be deduced from our general knowledge of the world’ (Tversky, 1977, p. 329). Libby and Luft (1993) help explain how repeated experiences affect knowledge development, noting that ‘ … the content and the organization of knowledge can be changed by decision makers’ learning opportunities’ (p. 428). Expertise develops through repeated experiences and the associated development of knowledge (Bedard & Chi, 1993; Nelson, Libby, & Bonner, 1995). And according to Craswell et al. (1995), ‘specialized industry knowledge is thus a component of auditor expertise in addition to the general knowledge base required for all audits’ (p. 301). Industry specialisation is associated with product differentiation and fee premia (Craswell et al., 1995).

4 As of 1 January 2005 all listed companies are required to follow the IFRS.

5 We employ the previous year's number of public clients in order to capture the public-client specialisation at the beginning of a company-year.

6 The audit report must be signed by at least one auditor. In 16% of the 420 observations, the audit report is signed by two engagement partners from the same audit firm. We use the value for the partner with greater specialisation when measuring public-client specialisation, and we use the value for the partner with greater experience when measuring partner experience. As a sensitivity test, however, we use the average of the partners’ public-client specialisation and inferences from these tests are essentially the same as those from the main analyses, although the significances are weaker. Based on interest from one reviewer, we also conducted analyses using the lowest number. We find that the significant results for our test variable become mainly insignificant using this specification. This makes sense because it seems most logical that the partner with greater public-client specialisation should be most important from a leadership perspective for the engagement; it also seems implausible that the effect of the partner with less public-client specialisation/experience would outweigh the effect of the partner with greater public-client specialisation.

7 We use the database provided by the Central Chamber of Commerce to identify the number of years of professional experience for each auditor. We acknowledge that it would be desirable to control for number of years auditing public clients, however, this information is not available.

8 The industry classifications are based on two-digit SIC codes. Zerni (2012) defines industry specialists as those with five clients in the same industry. Our sample is characterised by smaller industry groups and we observe that only about 13% of the engagement partners are classified as industry specialists. Our sample includes only 15 company-year observations audited by engagement partners that have more than two public clients in the same industry.

9 ‘The Finnish Corporate Governance Code’ recommends that the public companies establish an audit committee. Non-compliance with the recommendation must be identified and explained in the annual report.

10 For example, in one of our main estimations, column (1) of Table 5 Panel A, the highest VIF among these variables is 3.833.

11 Adapted from Zerni et al. (2012) the following logit model is used to estimate the probability of employing a public-client specialist audit partner: Prob(PUBS≥3) = a + β₁LOGASSETS + β₂LEVERAGE + β₃LOSS + β₄OCF + β₅STATEOWN + β₆AC + annual fixed effects + industry fixed effects + ϵ. STATEOWN equals one if the Finnish government has ownership in the company and zero otherwise. Industry fixed effects are based on one-digit SIC codes. Other variables are described in Table 2.

12 Consistent with our main variable definition, we observe that only the top 20% of partners have five public clients or more during t−1 and t−2. Consequently, PUBS≥5 represents the public-client specialisation in this analysis.

13 The variable PUBS≥10 represents the top 5% of partners based on the number of public clients during t−1 and t−2, and in this analysis they represent the partners with the highest level of public-client specialisation.

14 Results using cutoffs of 19, 18, 17, and 16 years yield the same inference.