Bell, T. B., M. Causholli and W. R. Knechel. “Audit Firm Tenure, Non-Audit Services, and Internal Assessments of Audit Quality.” Journal of Accounting Research, Vol. 53 No. 3, (June 2015): 461-509.
Does audit quality decline when audit firm tenure becomes long or fees from auditor-provided non-audit services become large?
The financial crisis of 2008 reignited a longstanding debate on the impact of audit firm tenure and auditor-provided non-audit services on audit quality.[i] This study investigates these concerns using data from an international audit firm’s internal reviews of audit process quality in 265 US audits of publicly and privately held clients. Primary analyses are based on two quality measures developed from the review data: 1) the total number of assessed audit deficiencies across 55 separate audit process activities; and 2) a composite assessment of the overall quality of the audit.
Summary of Results
Audit quality is lowest in first-year audits, improves shortly thereafter, and then declines somewhat as audit firm tenure becomes long. The decline in quality in the long-tenure range is attributable to audits of private clients. In public-client audits, quality increases slowly over the entire tenure range and is not significantly higher than in a first-year audit until the longest period where tenure exceeds 13 years. In contrast, quality in private-client audits improves quickly after the first year but declines with very long tenure to the point where it is indistinguishable from audit quality in the first year.
In tests of the impact of non-audit fees on audit quality, quality is not associated with non-audit fees in the full sample but, as above, the public- and private-client subsamples exhibit different patterns of association. The association of audit quality with non-audit fees is positive in audits of public clients and negative in the audits of private clients. For public clients, the probability of a high-quality audit is 7 percentage points higher for clients purchasing non-audit services than for clients not purchasing these services. Conversely, the probability of a high-quality audit of a private client purchasing non-audit services is 18 percentage points lower than for those not purchasing non-audit services.
Supplemental analyses show that audit hours are significantly higher in first-year audits and that audit partner specialization in the client’s industry is associated with higher audit quality in both the full sample and in first-year audits.
Implications and Limitations
In first-year audits, lower audit process quality and higher total audit hours are possible additional costs that should be considered in the ongoing debate on mandatory audit firm rotation. Moreover, study results are consistent with the notion that—even prior to the effective date of the Sarbanes-Oxley Act (SOX)—market and related regulatory forces disciplined auditors of public entities to achieve a high level of audit quality when tenure was long or fees from auditor-provided non-audit services were large. In order to serve the public interest, these considerations should be included in assessments of the economic costs and benefits of restrictions on audit firm tenure and non-audit services.
Furthermore, the results suggest that, in the private-client market, audit process quality declines in the long tenure range and when non-audit fees become large, which may be of interest to standard setters in the private sector (e.g., the Auditing Standards Board and US State Boards of Accountancy).
The study has several potential limitations. First, the data were collected in 2003, just after the de facto dissolution of Arthur Andersen (AA) and just before the effective date of SOX. AA was the predecessor auditor for 80 percent of the 120 first-year audits included in the sample. When tests are rerun without ex-AA audits, quality in the sample of first-year audits is no longer lower than quality in the remaining sample of post-first-year audits. If the unusual circumstances surrounding the ex-AA audits influenced the quality assessments, the finding of lower quality on a first-year audit may not be generalizable. Second, the data are from one period and pertain to a time characterized by an increased scrutiny of the audit process. This may result in evaluations that are more stringent than usual. However, this setting closely resembles the current US regulatory environment involving Public Company Accounting Oversight Board (PCAOB) inspections. Third, oversampling based on risk precludes meaningful extrapolation of the results to the firm’s full client portfolio. Fourth, the measures of audit quality are based on subjective judgments of individuals, which are subject to potential information processing limitations and biases. Fifth, to the extent that the processes are unique to the firm, results may not generalize to other settings.
Interested readers can download the full study free-of-charge by clicking on the paper title referenced above or here.
[i] See, e.g., European Commission, Green Paper: Audit Policy: Lessons from the Crisis, Brussels, Belgium (2010) and Public Company Accounting Oversight Board, Concept Release on Auditor Independence and Audit Firm Rotation, Washington, DC (2011).