Auditing is always about risk appraisal. That risk is always appraised at boundaries. That boundary might be where goods or services are acquired or sold by the entity being audited. It might be where legal status changes. It can be where the past becomes the future. But the principle never alters. Auditing determines the risk that a transaction is measured inappropriately when it changes state because of that transition across a boundary. And the problem is that basis on which group accounts are prepared requires that some boundaries are erased and others are artificially created - and this act of re-drawing is, we suggest, leading to outcomes that prejudice audit quality.
This requires explanation. We do not dispute that there are companies (and trusts, individuals, and other organisations) that control other entities with the intention that they trade with broadly common purpose whilst sharing some degree of central management and control. That is undeniable. But the problem with group accounting is actually inherent in this description of what a group is.
Firstly, the group is by definition not a single entity. It is necessarily made up of multiple entities or it would not be a group. Almost invariably each component element of that group is incorporated. And each of these entities will, almost invariably, have limited liability. Consequently each may well require audit in its own right. What that audit of the subsidiary will recognise is that the subsidiary entity is distinct, and has its own boundaries. It will also have its own management, governance structure, purpose and supposed independent legal requirement that it be managed in the interests of the shareholders (but with the implication that the interests of the creditors of the company are protected, at least when consideration of the appropriateness of the going concern concept is taken into account). Subsidiaries therefore have their own substance quite independent of the group of which they are a part and that substance has, then, to be recognised by their auditors.
However, those auditors are almost always from the same auditing firm as those who audit the group as a whole. But those auditing the group have to assume, because of the group accounting convention, that its subsidiaries have no boundaries and are instead just a bundle of transactions, some of which may be included in the group accounts, and others of which are not because they represent intra-group transactions. The separate entity that the auditor is, when considering the subsidiary, required to assume exists, is then simultaneously assumed to be of inconsequence by the very same auditor when considering the group. And transactions that have very real consequences for subsidiary companies have to be ignored at group level.
This Schrodingerian treatment of subsidiaries is, we think, the unseen and so far uncalled out quandary at the heart of modern auditing. That is because it exposes logical flaws, both from the viewpoint of those undertaking the audit and from the perspective of the user of the accounts, wherever they might be on the stakeholder spectrum.
This would not matter if it was just a technical issue, but there are real consequences to this failure. What it does is introduce boundary arbitrages into auditing as some costs appear within one set of boundaries but disappear in another. In other words, there are problems of incommensurability when the group does not appear to reflect the sum of its parts. This has been seen in practice: as the case of Interserve illustrates, the value of the subsidiaries as recorded on the parent company balance sheet looked very different to their value on the group balance sheet, after goodwill impairments were factored in. In that case the auditors appeared to take an entirely different view of the value of the subsidiary when preparing the group balance sheet from that they took when preparing the parent company balance sheet and yet they audited both and even published them in the same document. The two views appear to us to be irreconcilable. As a result audit quality is being compromised.
The second issue is that if audit is failing then this is because accountancy itself is failing. Those accounting failures include not making explicit the issues that arise as a consequence of geography (which can be made clear by country-by-country reporting); the consequences of transactions arising within the group (which could also be addressed by extending the scope of country-by-country reporting and reforming related party reporting); the scale of intra-group balances and reserves and whether those reserves might actually be distributable or not without accounting (not least with regard to tax) consequences arising; and the extent of assets available to protect the creditors of the group, not as a whole but in the entity against which they actually have a legal claim. All these are fundamental accounting weaknesses resulting from a failure to recognise the very real boundaries that do exist within groups, the need to appraise risks as transactions arise upon those boundaries and the obligation to report to suppliers of capital on the actual risks that they might face. This is a significant accounting and audit failure.
This then inevitably leads to the conclusion that current debate on audit failure is also missing the true explanation for it occurring. It is, of course, true that there are undoubted regulatory failings in the UK audit market and that audit quality could almost certainly be improved. We also think it is true that barriers between auditors and the providers of other services make sense. But addressing these issues will not solve auditing’s problems if accountancy does not recognise the risk in group accounts, because it does not recognise the boundaries where risks arise that are inherent within all groups but which are removed from view by group accounting. Only fundamental accounting reform that reforms group accounting so that the location of these risks is made clear will solve this problem. And until that happens auditors will continue to report on something that is little more than a work of fiction, which is what the group accounts of most companies currently are – they are a report on transactions that have never taken place in the way in which those accounts suggest that they have. Auditing fiction will never create reality, and that is why auditing cannot deliver what is expected of it at present.
Richard Murphy,FCA, is a professor at City University and Adam Leaver is a professor at the University of Sheffield