The concern is that requiring a building society to be a Public Interest Entity (PIE) has led to a lack of proportionality in building society audit regulation and oversight, which has driven up audit costs and reduced competition and choice. The BSA found that the audit cost for the smallest 11 societies is 9% of their profits, which certainly seems excessive and suggests something is wrong and needs to change. 

It is also somewhat anomalous that credit unions, in many ways akin to the smaller societies, are not PIEs but are subject to both a more lenient audit and prudential regulatory framework.  Likewise, it is odd that large investment firms that do not have listed securities are not PIEs yet may pose greater risks than building societies.   

EU law  defined PIEs as including credit institutions, such as building societies, and set out various requirements. Two reasons were given for the requirements: “the smooth functioning of the internal market, while achieving a high level of consumer and investor protection”. It could be argued the first of these two reasons no longer applies post Brexit, and so the legislative requirements could be relaxed to some degree.

The requirements look to enhance audit quality and affect both individual audits and the PIE audit firm.  For example, the PIE auditor is prohibited from providing specified non-audit services (e.g. tax), is required to undertake an independent engagement quality control review for each PIE audit, faces mandatory audit rotation, and then might find the audit reviewed by the regulator.  PIE audit firms must publish a Transparency Report annually, and the FRC exercises regulatory oversight. 

Issues

What is the real audit cost?

If societies were no longer PIEs this would allow some of the legislative requirements to be removed.  Or, as the application of auditing standards requires the use of professional judgment, relax some of the FRC’s expectations for a good audit to reduce costs. So, the auditor could apply fewer procedures, gather fewer forms of evidence, reduce training and administrative procedures such as documentation levels.  It is not clear, however, to what extent the requirements should be removed and the expectations relaxed. It depends on the extent to which the FRC’s expectations “gold plate” and go beyond what is necessary for a good society audit. 

It could be argued that audit costs have merely caught up with the real cost of providing a good audit generally. Auditing standards (ISAs) and practice generally have improved (appropriately so) since the 2007/08 Global Financial Crisis to drive up audit quality.  As the BSA report notes, the audit cost for two credit union audits likewise increased significantly over a five-year period. By curtailing non-audit services and fees, the legislation is also potentially revealing a cross-subsidy that existed.
It should also be noted that audit work will have increased as financial statements have grown ever longer, and balance sheets have become more complex as accounting standards have evolved. Even if all information within the annual report is not audited it still has to be reviewed for consistency with the audited information. While not affecting all societies, typically the smaller societies that have stayed on UK GAAP, but international financial reporting standards replaced incurred credit losses with expected credit losses - an inherently more judgemental and risky approach that needs more audit work.

Even if societies ceased to be PIEs, the same auditing standards would be in force for which auditors have already developed approaches. It may not be cost effective or desirable to change these for a new category of non-PIE audits, especially if the building society business model is unchanged. There would still be an audit regulator (though not the FRC), and the market may have the same expectations for a good audit.  Moreover, the FRC’s expectations are likely to leech into the market’s view of acceptable approaches to non-PIE audits where business models are similar.  

If a good audit is simply too costly, the implication is that the business model for the smallest societies may no longer be viable, without a conscious decision that the audit standard should be reduced, which may affect the PRA’s approach to regulating building societies and increase the risk to the Financial Services Compensation Scheme.   

What are the audit market barriers?

Removing PIE status would certainly open the market to more and smaller audit firms, notably those that audit credit unions, thereby potentially introducing more competition and choice. If these audit firms are unencumbered by PIE requirements or expectations, they may be able to conduct a cheaper audit. The short run costs may however be similar, as current audit fees provide a benchmark and new auditors incur higher first year audit costs. The risk of removing PIE status, however, is that audits are weaker and might fail to meet an adequate standard.   

But there are 30 PIE auditors . Of these, only the 6 Tier 1 audit firms undertake society audits: and almost all banking audits.  This suggests that PIE status alone is not a barrier to market capacity.  There are enough PIE players: they choose to not compete for society audits.  

Auding can be a challenge whatever the industry, but the barrier is perhaps that building society audits are still seen towards the riskier end, as much of the balance sheet is an estimate (eg credit losses or fair value on derivatives) while the sector is highly regulated. Smaller entities may also have less developed controls upon which an auditor can place reliance. A good audit that is higher risk is more costly to deliver and requires a commensurate higher reward: if the sector cannot bear or is not willing to bear this cost, new entrants will not enter the market.    

To enter the society / bank audit market also requires significant investment, but firms face a chicken and egg question – whether to invest and then look to win business or win business and then invest to build the necessary capability.  Either route carries risk: the audit firm does not gain critical mass in audits won to compensate the upfront investment; while the latter runs the risk of falling “fowl” of the audit and prudential regulators (FRC ad PRA) if there is insufficient investment to deliver a good audit. Increased audit rotation also brings the risk that the auditor is moved on before first year investment costs are fully recovered. 

Another issue is that smaller societies are regional. For some, this might mean there is really no choice beyond the existing large audit firms. 
A further point is that it can be more lucrative to provide non-audit services (eg internal audit), as these are more likely to achieve a higher recovery rate than audit and are without the same regulatory risk.  Even if PIE requirements did not apply, general auditor ethics and independence standards might mean this trade-off is still to some extent made.  

Finally, a single bank / society failure coupled with a bad audit may be sufficient to damage an auditor’s reputation and drive them out of the market.  They may judge the risk is too high to consider entering the market. 

Impact on the PRA

The PRAs’ supervisory approach is based on a good-quality audit being undertaken. If societies are no longer PIEs and their audits are undertaken by smaller firms outside the scope of FRC oversight, there is a risk that the PRA considers the audit no longer fully meets its needs. As a result it may take additional measures, such as commissioning skilled persons reports, which might lead to an increase in regulatory costs. 

So what to do?

An audit costing 9% of profits suggests something needs to change.  While there is some merit in removing the PIE status for certain societies, it may be simplistic to suggest this is all that is required.  It seems unlikely that removing PIE status would suddenly lead to a significant fall in costs, and it is not without risk (eg weaker audits). There is also the challenge where to draw the boundary between societies that are PIEs and those that are not.  

There may be some actions that societies and their audit committees should explore to reduce costs. For example, as most societies (as with banks) have a December year-end, they could explore staggering their year-ends to coincide with potentially quieter audit periods.  Or there may be actions that could facilitate the audit – such as enhancing controls or audit trails – which could help reduce audit costs. Planning well ahead of an audit tender may help identify more potential bidders, and actions that could encourage them to bid.   

The BSA, possibly with help from ICAEW Financial Services Faculty, could undertake further work to better understand the cost drivers for society audits and explore whether there are ways and means to bring them down.   

The BSA/Faculty could also consider whether there are gaps in the training material and resources they provide to their members, that might assist with upskilling new audit entrants and lessen the burden of any initial investment needed.      

More imaginatively, and requiring legislative changes, longer-term options might be to explore whether specific features of the PIE and audit regimes could be changed in certain circumstances. For example, the frequency of the audit could be reduced – so every two or three years rather than annual – with an interim review in the intervening years or an audit (eg ISA 805) or assurance review of certain key numbers (eg of the credit provision only). The lack of shareholders and being subject to FCA and PRA oversight might justify fewer audits.  

In time, advances in technology - data collection, processing and analytics – will enable new audit techniques and may reduce costs. While again there is an upfront cost that needs to be covered, that is likely to be spread across a much wider population than the societies.