Cost and Informativeness of Regulatory Reports: Evidence from the UK

Abstract

The 2006 EU Directive established an EU-wide system for public oversight of the audit profession. In the UK, since 2008 the Financial Reporting Council (FRC) has published inspection reports for major audit firms, which include overall quality ratings for the individual audit engagements of each audit firm under review. This study examines the FRC’s ratings, and measures their impact on audit fees and audit firm switching. A significant increase in audit fees is found when the audit firm has a higher proportion of engagements with deficient ratings, probably arising from the additional effort and resources needed to meet the FRC’s requirements. This impact is more concentrated among clients with Big 4 audit firms. However, there is no evidence that FRC ratings affect clients’ likelihood of switching audit firms, suggesting that inspection results may not signal audit quality, and thus do not affect clients’ audit firm appointment decisions. The results provide evidence that inspection ratings may increase audit costs, but may not be valuable in distinguishing audit quality, and thus have no effect on audit committees’ audit firm appointment decisions. This finding advances understanding of the effectiveness of the audit inspection regime, and provides auditing regulators with guidance on policy making.

Keywords:

• inspection ratings
• audit fees
• audit firm switch

JEL Classifications:

• M42
• M48

Acknowledgements

I thank Annita Florou, Peter Pope, Mark Clatworthy, Kevin McMeeking, Amedeo Pugliese, Collin Clubb, as well as seminar participants at the Emerging Researcher Consortium, 21th Annual Financial Reporting and Business Communication Conference, King’s College London and University of Nottingham Ningbo China for their comments.

Disclosure Statement

No potential conflict of interest was reported by the authors.

Notes

1 According to the FRC (2016), these clients are ‘related to FTSE 100, FTSE 250, other listed and other major public interest entities’.

2 Other recent papers that analyse enforcement activities include Christensen et al. (2017) and Florou et al. (2018). Christensen et al. (2017) examine the effect of proactive financial reporting enforcement intensity on shareholder wealth. They find that increased enforcement intensity is associated with greater disclosure, more resources expended on regulatory compliance and higher stock-market liquidity. Nevertheless, the market reaction for targeted industries is negative, and firms with stronger pre-existing private oversight experience the greatest drop in equity values. Florou et al. (2018) examine the costs and benefits of proactive financial reporting enforcement. They find that increased enforcement relates to increased audit fees and more conservative accounting accruals. Specifically, the increase in audit fees is more concentrated among companies in the AIM market, while accruals are conservative in the main market.

3 According to Gunny and Zhang (2013), a report is ‘clean’ if no deficiencies are identified, ‘deficient’ if one or more audit deficiencies are found, and ‘seriously deficient’ if the deficiency relates to a ‘failure to identify a departure from GAAP’ and/or a particular deficiency results in a ‘restatement’ of the financial statements.

4 According to Acito et al. (2017), relative exposure to deficient auditing is measured as the difference between deficient auditing exposure with the current audit firm and the adjusted audit firm deficiency exposure that the client would face with other audit firms. Specifically, by matching a keyword list (based on deficiencies identified in PCAOB inspection reports) with the number of times these keywords relate to each standard identified in the client’s 10-K filing, audit firm deficiency exposure is calculated as the client’s standardised keyword count for an accounting standard multiplied by an indicator variable which is equal to one if the PCAOB identifies a client’s audit firm as having a deficiency relating to that standard.

5 Checks for previous studies (e.g Osma et al., 2017) and online searches reveal that Spain and the UK are the few countries that disclose firm-specific details on disciplinary sanctions in Europe. Sweden is the other country where inspection reports are made publicly available, but these are disclosed only for some Big 4 audit firms.

6 The PCAOB was established in 2002 via the Sarbanes-Oxley Act (SOX). All public accounting firms (i.e. audit firms) that audit SEC-registered companies are inspected by the PCAOB. The PCAOB has the right to undertake special investigations and take disciplinary measures when necessary, whereas in the UK investigations and disciplinary measures are undertaken by the AADB, a separate institution from the FRC (IFIAR, 2013). Neither the PCAOB nor the FRC disclose clients’ names in their inspection reports.

7 Checks with FRC staff confirmed that companies listed on AIM can only be selected if they meet the FRC’s threshold (e.g. a market capitalisation of £100 million). Only a minority of AIM companies meet this threshold, so most are not within the scope of FRC inspections.

8 In most cases, the FRC gives fiscal year ends rounded to a month. Hence, when forming test variables, the last day of the month is used as the lower and upper limit. For example, when the fiscal year end is between 06/2008 and 06/2009, it is treated as 30/06/2008–30/06/2009. The results were robust to using the first day of the month as the lower and upper limit.

9 According to the FRC (2014), one December 2011 year-end engagement for E&Y was reviewed during the 2013/14 inspection, and thus belongs to the previous inspection period. When forming test variables, this engagement was excluded and the fiscal year end covered in 2014 was redefined as 04/2012–03/2013. The results were robust when all E&Y’s firm-year observations for 2014 were dropped.

10 In the most recent inspection reports, fiscal year ends are not given in annual reports, but the FRC confirmed to the author that the periods were similar to the previous inspection report, generally from June 2014 to April 2015.

11 In alternative models, industry-clustered standard errors were used when calculating t statistics for all analyses.

12 To mitigate cross-sectional variability arising from Worldscope audit fee data potentially including non-audit fees paid to audit firms, robustness tests were conducted using changes to the variables in the regressions. These produced results in line with the main results.

13 For the audit fee and audit firm switching analyses, the fiscal period for the most recent inspection ended in April 2015. In order to measure the impact of lagged ratings, data for 2016 could only be used up to April, resulting in fewer observations for 2016 than for other years.

14 The only exception was ROA, which was positively correlated with CFO and negatively correlated with LOSS. The empirical findings remained qualitatively unchanged after dropping ROA from the audit fee analyses.

15 Although audit firm-fixed effects are included in the baseline model, to check for differences between Big 4 and non-Big 4 audit firms, a Big4 dummy was included in the baseline model, and the results did not qualitatively change.

16 The Code also requires audit committees to provide shareholders with information on how they have carried out their responsibilities, including assessing the effectiveness of the external audit process (FRC, 2012).

17 In the regression for the non-Big 4 subgroup, the OPINION variable was omitted since its value is always ‘0’ for all firm-years.

18 OLS regression (with client-fixed effects) and probit models were also run for switching analysis, and the results did not change qualitatively.

19 Similarly to Weber et al.’s (2008) study, which was based on a unique German setting, most control variables in the audit firm switching analysis were insignificant, probably because the UK and Germany have fewer observations than US studies, so a relatively low frequency of audit firm switching may weaken the power of the test.

20 Some observations in the audit firm switching analysis were dropped when running the regression because some audit firm indicators perfectly predict the dependent dummy variable. As a sensitivity test, audit firm-fixed effects were dropped from the baseline model, and the results did not change qualitatively.

21 Similarly to the audit fees analysis, a BIG4 dummy was included in the switching model, and the results did not qualitatively change.

22 The test for lagged inspections performed in the audit fees analysis was not included because the construction of the lagged rating variable reduces variation in SWITCH. After creating the lagged variable, only one firm-year observation had switched audit firms between years.

Additional information

Funding

This work was supported by University of Nottingham Ningbo China: [grant number I01191000053].