4Separating audit from non-audit
The case for separation
97.The CMA’s update paper on its market study, published in December 2018, set out a number of preferred remedies, including a split between the audit and advisory business.179 The CMA is clear that auditors should focus exclusively on audit to secure higher quality, and not also on selling consulting services. Separation is one way to achieve this aim.180 In this chapter, we explain how separation would improve quality by:
- creating a strong audit culture in the firm and eliminating tensions with the very different culture of advisory services;
- enhancing transparency; and
- making audit truly independent by ending the subsidies from the rest of the firm.
98.Separation could also improve choice by removing barriers to competition and reducing conflicts, depending on the model of separation adopted. We discuss the forms that separation can take later in this chapter.
99.To a greater or lesser degree, the Big Four acknowledge the benefits of separation. KPMG’s response to the CMA’s update paper recognises that audit is too opaque and needs more transparency.181 In oral evidence, the head of KPMG, Bill Michael, expressed support for separation:
[O]ur business model [ … ] needs to become more transparent because it is impairing confidence in audit quality. We have heard before that there is scepticism as to whether large multidisciplinary firms are focused enough on audit, so some form of separation is the right direction of travel and we have started doing that.182
100.Deloitte agreed that separation can address “cultural and incentive concerns” and “serve to increase audit quality”.183 Their response to the CMA explains that separate governance and management can achieve “significant and far-reaching enhancements” to the audit business, such as:
an overriding responsibility to act in the public interest, an overarching performance management structure aligned solely with the delivery of audit quality and management process and systems to reinforce that.184
Separation would demonstrate “a culture of quality, independence and objectivity”, and eliminate “undue influence from the wider (non-audit) business”.185
101.BDO’s Scott Knight argued that the audit practice needs to be governed by auditors, that separate financial statements must be available to the regulator and that the issue of cross-subsidies between audit and non-audit should be dealt with.186
Audit culture
102.PwC emphasised the importance of creating the right culture in an audit firm:
We agree with the CMA that promoting the right culture within an audit practice is critical; we are open to remedies designed to enhance audit firm governance in order to better embed the right culture and increase the focus on audit quality.
103.We believe that a strong audit culture of independence and challenge is one of the most important, if hard to quantify, determinants of audit quality. Culture is crucial because “auditing, by its very nature, is judgemental and based on human decisions and actions”, as argued by the FRC.187 We are therefore concerned about audit representing a small and shrinking part of the multidisciplinary firms. The smaller the audit practice becomes as a part of the firm, the less we would expect it to shape the culture, priorities and investments of the firm. Firms argue that audit is at the core of their identity (their “raison d’etre”).188 We are sceptical that this can be true, when non-audit is already four times larger (and growing) than audit across the Big Four.189 As the relative size of audit declines, there is inevitably a point at which audit can no longer be regarded as core to the firm. It is perhaps a sign of the status of audit in these firms that none of the Big Four in the UK is headed by an auditor. Indeed, in their evidence to us, the non-audit heads of these multidisciplinary firms revealed an at times shocking lack of knowledge about the purpose and nature of audit.190
104.Secondly, since the culture of advisory (helping management) is by nature antithetical to that of audit (challenging management), a multidisciplinary firm needs to have a strong and distinct subculture in its audit practice. However, evidence suggests that this is hard to achieve. The CMA reported a recent study on the Big Four in the US that found a negative relation between the importance of non-audit services at the firm level and audit quality.191 The study concludes that the observed impact on quality can only be explained as the result of internal competition for resources, non-audit fee pressure on audit partners, reduced management attention to audit, or a change in the firms’ culture. The CMA also cites other studies that indicate how the culture of “client service” within accounting firms may conflict with audit notions of independence and public service.192
105.The FRC’s Audit Culture Thematic Review (May 2018) offers convincing evidence that predominantly non-audit firms have a culture problem. The FRC found:
- Culture (being purpose, values and encouraged behaviours) is designed for the whole multidisciplinary firm.
- In some firms, audit specific values such as objectivity and independence are not sufficiently prominent.
- All firms could do more to promote to partners and staff the purpose of an audit and the societal value that it brings.
- Only in one firm was there a clear sense of audit being there to provide confidence in the capital markets.
- Only four of the eight firms prominently included improving audit quality within their whole-firm strategies.
- Most firms recognised that their multidisciplinary nature contributes to cultural challenges, but few firms have clearly articulated those challenges or taken specific, explicit actions to address them. Examples include:
- The tensions that can be created by audit sometimes being less profitable than other lines of service.
- The extent to which the multidisciplinary firm model contributes to a mindset of partners and staff that is focussed on providing the best possible client service, with the client defined as the company to which the service (including audit) is being provided. Such a mindset may make it more difficult for auditors to apply an appropriate degree of professional scepticism and challenge in their work.193
106.As to increasing choice, the CMA’s update paper explains that separation “might allow relaxation of the conflicts rules and therefore a return to greater choice among the Big Four”.194 This depends on the degree of separation. Relaxing conflict rules might not be possible or desirable without full legal separation of audit from non-audit, but full separation is costly.
Degrees of separation
107.Three alternative models of separation are discussed in this debate:
- governance separation,
- operational separation (to separate economic interests), and
- structural separation (to create two fully independent entities).
108.Governance separation is the shallowest and easiest to implement. Structural separation is the fullest and most disruptive. Operational separation is a compromise. A fuller separation delivers more benefits and with greater certainty, but also entails higher costs and risks. We discuss below the balance of risks and rewards of the different degrees of separation.
Governance separation
109.A governance separation would be an internal arrangement to separate the running of the audit practice from the rest of the firm. Everything else would still be shared—including profits. Precise models might differ, but Deloitte’s proposed version includes the following features:
- An audit business with its own CEO, overseen by a new separate governance board and a completely independent Chairman;
- The Audit Governance Board [ … ] would have clear terms of reference including maintaining a culture of audit quality within the audit practice;
- Oversight and enforcement of clear, transparent policies that link partner and staff remuneration to audit quality, overseen by the governance board;
- A firm-wide strategy and budget set at a firm level, with the level of investment in audit agreed by the Audit Governance Board, to ensure investment levels are consistent with its aims of safeguarding audit quality;
- Disclosures akin to a listed company together with an annual report and an AGM.195
110.A governance separation would deliver some improvements in transparency and focus on audit quality, but it does not go nearly far enough in tackling the conflicted commercial culture of the firms. We share the CMA’s concerns in the Audit update paper that auditors are being damagingly exposed to the commercial pressures and culture of non-audit services in multidisciplinary firms. This culture undermines the audit mindset of challenge and professional scepticism.196 A clear and tangible way in which wider business interests seep into the audit practice and compromise its culture is that the vast majority of profits distributed to audit partners come from the non-audit business (see next paragraph). A fatal shortcoming of the governance split is that it does not in any way separate the economic interests of the audit and non-audit arms of the firms. We agree with the CMA that “changes at the firm level are [ … ] necessary to ensure a single-minded focus on audit quality”.197
111.Across the Big Four as a whole, non-audit services accounted for 79 per cent of total revenues in 2018.198 In other words, non-audit fees are four times larger than audit fees. As non-audit has been consistently more profitable than audit,199 the percentage of the firms’ profits accounted for by non-audit is even greater. Non-audit profits feed directly into the remuneration of audit partners, who receive a share of the overall profits of the firm. In the words of the CMA, “[t]his means that audit partners are incentivised to care about the overall performance of the firm (the majority of which relates to non-audit services)”.200
Cross-subsidy
112.The Big Four reported overall profits from their audit practices.201 The fact that audit appears profitable, however, does not mean that it is not subsidised. On the contrary, hidden subsidies make audit more profitable than it would otherwise be. Without subsidies, audit would look a lot less profitable or may not be profitable at all based on current prices. Below we look at three forms of subsidy currently taking place, none of which would be ended by a governance split.
113.The remuneration of audit partners is effectively subsidised by non-audit profits. KPMG’s response to the CMA’s update paper explains that without non-audit profits, “audit partner remuneration could not be maintained at a sufficiently competitive level to continue to attract the requisite talent based on the current audit prices.”202 We do not think that this is a healthy state of affairs. Audit partners must not be or feel indebted to non-audit partners. That frame of mind can only lead to audit partners being mindful of the interests of the non-audit practice, when we expect them to serve the interests of the firm, its shareholders and the wider public.
114.A second form of subsidy from non-audit to audit stems from the specialist skills of non-audit staff being made available to audit at cost, rather than at an arm’s-length market rate.203 A third implicit subsidy is that audit does not bear its fair share of certain costs such as pensions and professional indemnity insurance.204 In light of this list of subsidies, we conclude that the true cost of audit is not currently reflected in the audit fee. By implication, audit is also less profitable than might appear from reported figures.
115.We question why non-audit partners are apparently content to provide this subsidy. EY argued that the audit practice deserves compensation for making “an important contribution to brand and reputation, and to the firm’s risk management and compliance frameworks”.205 From that perspective, audit is a branding investment which helps the non-audit practice win business. Audit also gives firms special access to corporations and executives (for example via contacts and networking), and that access may be worth paying for because it helps secure non-audit contracts. This underlines our argument over conflicting interests.
116.Far from being an objection, as the firms argue, ending cross-subsidies and letting prices rise if need be is a desirable development. In oral evidence, Sir John Kingman stated that this was a key reason for the sort of separation proposed by the CMA: “An important attraction of what the CMA is proposing is that there would be a clean ring-fencing within the firms, such that cross-subsidy could not occur.”206
117.Finally, the scale of subsidy currently enjoyed by audit in multidisciplinary firms makes it all but impossible for an audit-only model to be viable. It only serves to reinforce the dominance of multidisciplinary firms.207 That runs contrary to the goal of improved choice.
118.The opaque economics of audit undermine independence, erode trust and stifle competition. Audit can only be transparent and independent when it is fully priced. It will only be fully priced when it is no longer subsidised. Therefore, subsidies must end and audit must stand on its own two feet. We conclude that governance separation does not go far enough on the grounds that it fails to deliver independence and does not end cross-subsidies.
Operational separation
119.Operational separation goes one step further than governance separation by also separating economic interests. The CMA proposed the following key features for the operational split:
a) separation of the audit and non-audit businesses, with the audit business having a separate board, chief executive, staff and assets;
b) separate profit and pension pools for the audit and non-audit entities;
c) restrictions on audit partners (but not staff) moving between audit and non-audit businesses of the same firm;
d) transfer pricing arrangements between the two entities, for example to support use of non-audit staff on audits; and
e) both the audit and non-audit businesses could share some central operations, systems, branding and know-how, and both would remain part of the same multidisciplinary network.208
120.It is (b) and (d) that deliver the separation of economic interests, over and above governance separation: the operational split is governance separation plus separate profit pools and proper transfer pricing between audit and non-audit to end cross-subsidies. These extra measures have the potential to deliver independence and transparency needed, if effectively implemented and carefully monitored.
121.The main objection from the firms to separating profit pools is that this might lead to higher audit prices.209 However, as we argued above, letting audit be fully priced is precisely the goal that we are pursuing for competition and transparency reasons. Independence is worth paying for.
122.An economic separation of audit and non-audit is highly desirable. We recommend that the CMA at the very least implements the proposed operational split to achieve the separation of economic interests.
Structural separation
123.The CMA’s “structural split” is the most radical proposal. It involves splitting audit and non-audit into two separate and independent legal entities. The proposed structural split includes the following features:
- any audit firm over a certain size threshold would be prohibited from providing non-audit services in the UK;
- the UK audit firm would then become a stand-alone firm, but could remain part of a multidisciplinary international network;
- audit-only firms would need to replicate in-house the non-audit expertise required to support an audit (for example, by recruiting non-audit specialists to work on audit engagements), or contract externally from an independent non-audit practice.210
124.We believe that there is a strong case for independent audit firms. A full, clean legal separation would comprehensively address the real and perceived problems we have identified above: cultural tensions, conflicts and concerns over the financial independence of audit services. It would be far more likely to achieve the benefits of separation than an operational split: it would be fully transparent, more likely to foster trust and simpler to monitor. It could also obviate the need for some of the other independence remedies such as cooling off periods, and corporate hospitality could simply be banned for the audit-only firms.
125.Full legal separation could also increase choice, because it would reduce the conflicts between audit and non-audit that can disqualify firms from auditing companies they supply other services to. Currently, there is a 70 per cent cap on non-audit services to audit clients that applies to the firm, but not to the other members of the network. However, the ‘blacklist’ of services that cannot be provided to the audit client applies to all the members of the network, and so would continue to limit choice even after a structural split.211
126.The benefits of legal separation are compelling. However, we recognise that a structural split involves significantly greater costs and risks than an operational split.
127.The Big Four argue that a structural break-up would lower audit quality for the following reasons:
- Smaller audit firms would be more dependent on their large clients, and so less able to challenge them;
- Audit-only firms would find it harder to recruit and retain talent;
- Audit-only firms would not benefit from the vast range of specialist skills and technologies available to multidisciplinary firms.212
We tackle these objections in turn.
Greater dependence on large clients
128.We acknowledge that smaller firms would be more dependent on large clients. Nonetheless, whether this greater dependence affects auditor behaviour depends on the leverage that company management can exercise against the auditor. In this context, the main threat that management can exercise is firing or not reappointing the auditor. While this is no small threat, we are confident that our package of recommendations, together with Sir John Kingman’s new regulator (ARGA) and the CMA’s proposals, can minimise this risk, for the following reasons:
- We are recommending fixed-term, non-renewable contracts of seven years, so that auditors need not worry about being reappointed or fired for doing their job (see Increasing the Frequency of Audit Rotations);
- The CMA’s recommendation for regulatory scrutiny of audit committees would ensure that auditors are selected based on quality and willingness to challenge management;
- ARGA would have the right to appoint the auditor where there are serious concerns with the audit. Where the auditor resigns or is fired before the end of their term, ARGA would automatically investigate.
Recruitment and retention
129.We do not find the recruitment and retention objection convincing. The CMA looked at this issue in the Audit update paper and found:
[ … ] evidence [ … ] submitted to us, ranging from 2011 to 2018, suggests that the number of audit staff that permanently moved into or seconded into a non-audit team within their firms is relatively low compared to the total number of either audit staff or graduates.213
130.Moreover, we reject the notion that a structural split would stop audit staff from moving to non-audit at some point in their career. Part of the attraction of working for firms that are part of a global network is that they already can and do offer opportunities to move to a member firm in another country. If the non-audit business is allowed to remain part of the network (a decision for the CMA), a structural split would make it no more difficult to move from audit to non-audit than it is to move from the UK firm to the US or Australian firm.
Specialist skills and technologies
131.The third objection is stronger and comprises two parts: specialist skills and specialist technologies. The resourcing of specialists would no doubt be costlier and less seamless after legal separation, but we explained why this is desirable when discussing the merits of economic separation. Moreover, developing some of these skills ‘in house’ would ensure that these specialists share the culture, values and goals of audit. For those skills that the audit practice is unable to deliver in house, the CMA expects the practice to be able to contract for those externally. Currently, audit firms rely on expert input of non-audit specialists for between 10 per cent and 20 per cent of a FTSE 350 audit: not insignificant but by no means prohibitive.214 In time, the potential additional cost of contracting for these services should create a positive incentive to develop as many of those skills as possible in house.
132.On technology, we are unconvinced that legal separation would prevent the audit practice from accessing and/or investing in the technology it needs. The Big Four already share a large number of key technologies across their networks.215 As a member of the network, we see no reason why the audit-only firm would be stopped from accessing these technologies, or from investing in new technology together with the UK non-audit firm or with the US practice. There is nothing in a structural split to stop firms from investing together. Indeed, that is one of the key advantages of being a member of a large network.
133.In both the cases (skills and technology), deep and extensive collaboration is already taking place between legally separate and independent members of the network.216 Every time a multinational is audited, the Big Four provide an integrated international service which relies on foreign auditors auditing the subsidiaries based in their countries. Legal separation is no barrier to providing a seamless international audit217 and different examples of varying types and degree already exist.218
134.Some countries’ ownership rules, designed to safeguard the independence of auditors, in effect separate audit from non-audit. One example is France, where 75 per cent of the capital and shareholding of an audit firm must be held by statutory auditors. 75 per cent of the management, administrative and supervisory bodies must also be statutory auditors.219 The effect of these rules is described in the Transparency Report of PwC France, which explains that audit, advisory and tax services are delivered by separate legal entities. Audit services are “grouped in the holding company called PwC Audit, owned by natural persons who for the most part are professional auditors and/or chartered accountants, and who finance the company.”220
135.We note that similar independent ownership requirements applied in the UK with respect to law firms until 2012.221 Well before then, the Big Four still considered it worthwhile to set up independent, separately owned law firms. To give one example, PwC’s legal arm operated as a separate and independently owned legal entity up until 2016.222
136.It is clear that there are well-functioning models of legal separation and audit-only firms.
Conclusions on separation
137.We are persuaded that there are models of legal separation that are both preferable to the status quo and perfectly feasible. The precise model should be for the CMA to decide, including whether the fully separated non-audit business should or should not be allowed to remain part of the global network.
138.We have also set out reasons to be highly sceptical of the objections advanced by the Big Four. We do recognise the validity of one remaining objection: legal separation would be complex, disruptive and costly to implement. The CMA will have to weigh these costs against the benefits. That cost-benefit analysis should take into account the fact that implementation costs are one-off, whereas the benefits of separation will recur every year for the long term; and the fact that three of the Big Four sold their consultancy practices between 2000 and 2002,223 without apparent damage to the business.
139.We agree with the CMA that “objections to full separation are overstated”. We found the objections against full legal separation to be very weak, as membership of the global network provides an effective avenue to pool resources, access staff and share technologies. We also found well-functioning examples of legal separation and audit-only firms.
140.On the other side of the equation, legal separation offers benefits on multiple fronts: quality, independence, culture, transparency, trust and to some extent, choice. These benefits of separation are large, and in our judgement, worth incurring significant costs.
141.We encourage the CMA to aim for full legal separation of audit and non-audit services. The CMA should look to the long term, and not let one-off, short-term implementation costs weigh too much in its calculations. If the operational split is chosen instead, the CMA and ARGA should conduct a review of the arrangements after three years to determine whether the split has ended cross-subsidies and improved culture, independence and transparency. If not, we recommend that the CMA then move to implement a full structural break-up of the Big Four into audit and non-audit businesses in the UK.
179 CMA, CMA proposes reforms to improve competition in audit sector, (18 December 2018).
180 CMA, CMA proposes reforms to improve competition in audit sector, (18 December 2018).
181 KPMG, Response to CMA’s Audit update paper, (25 January 2019), p 3.
183 Deloitte, Response to statutory audit market study, (21 January 2019), para 9.1.
184 Deloitte, Response to statutory audit market study, (21 January 2019), para 7.6.
185 Deloitte, Response to statutory audit market study, (21 January 2019), para 7.7.
187 FRC, Audit Culture Thematic Review, (May 2018), p5.
188 EY, Response to Update paper: Statutory audit market, (25 January 2019), Appendix A p 2.
189 CMA, Audit update paper, (18 December 2018), p 28.
190 Oral evidence of 30 January 2019.
191 Meckfessel, M. D., and D. Sellers (2017). The impact of Big 4 consulting on audit reporting lag and restatements. Managerial Auditing Journal, 32(1), 19–49.
192 See Anderson-Gough, F., C. Grey, and K. Robson 2000. In the name of the client: The service ethic in two professional services firms, Human Relations, 53(9), 1151–1174.
193 FRC, Audit Culture Thematic Review, (May 2018), pp 7–9, 16, 29.
194 CMA, Audit update paper, (18 December 2018), para 4.114.
195 Deloitte, Response to statutory audit market study, (21 January 2019), para 7.7.
196 CMA, Audit update paper, (18 December 2018), p 8.
197 CMA, Audit update paper, (18 December 2018), p 8.
198 CMA, Audit update paper, (18 December 2018), p 28.
199 CMA, Audit update paper, (18 December 2018), p 76.
200 CMA, Audit update paper, (18 December 2018), p 83.
201 CMA, Audit update paper, (18 December 2018), p 83.
202 KPMG, Response to CMA’s Audit update paper, (25 January 2019), para 21.15.
203 EY, Response to CMA’s Audit update paper, (25 January 2019), p 4.
204 EY, Response to CMA’s Audit update paper, (25 January 2019), pp 2–3.
205 EY, Response to CMA’s Audit update paper, (25 January 2019), Appendix A, Part 5.
207 The point that cross-subsidies make it difficult for a stand-alone audit firm to compete with the large multidisciplinary firms is also made by the CMA in the Audit update paper, (18 December 2018), p 79.
208 CMA, Audit update paper, (18 December 2018), pp 117–8.
209 This is a logical consequence if audit currently enjoys subsidies from non-audit, as explained in the previous section.
210 CMA, Audit update paper, (18 December 2018), p 117.
211 EUR-Lex, Audit Regulation, Article 4(2).
212 For example: Deloitte’s and PwC’s responses in: Carillion: Responses from Interested Parties to the Joint Report, 12 July 2018, pp 22–3, 34–5; EY, Response to CMA’s Audit update paper, (25 January 2019), Appendix A, Part 5; KPMG, Response to CMA’s Audit update paper, (25 January 2019), p 45–8.
213 CMA, Audit update paper, (18 December 2018), p 121.
214 CMA, Audit update paper, (18 December 2018), para 4.128.
215 See the letters from the Big Four to BEIS Committee Chair in answer to the questions posed in the Chair’s letter dated 13th February 2019.
216 Oxera, Ownership rules of audit firms and their consequences for audit market concentration, October 2007, pp 103–4.
217 On the contrary, “one of the primary objectives of the network’s legal structure is to coordinate business development, skills and capabilities development, and the management of liability risk”. In: Oxera, Ownership rules of audit firms and their consequences for audit market concentration, (October 2007), p 104.
218 The above Oxera report explains how US accountancy firm McGladrey & Pullen LLP was separated between audit and non-audit after being acquired by H&R Block. Post separation, the audit business was owned and managed by audit partners, independently of the rest.
219 Oxera, Ownership rules of audit firms and their consequences for audit market concentration, October 2007, p 36.
220 PwC France, Rapport de transparence, (October 2018), p 15 (our translation).
221 ICAEW, Will accountants embrace multi-disciplinary practices?, (5 November 2014).
222 Lawyer Monthly, PwC Legal – An Overview, (December 2018).
223 Accountancy Age, How the Big Four have returned to consulting with a bang, (September 2015).
Published: 2 April 2019