Panel on Tax, Audit and Accounting

Written evidence submitted by Ernst and Young (SH 033)

1.1 How, if at all, does the tax system encourage leverage in banks?

1. We do not consider that the tax treatment of debt and equity has a significant effect on the amount of leverage undertaken by banks. In our experience, the proportion of equity and debt in the banks is primarily driven by commercial (non-tax) considerations and the restrictions imposed by regulation over the amount and type of capital held.

2. The tax deductibility of interest is a relevant factor for banks in determining the true cost of funding, but is rarely the main driver in the form of funding undertaken.

3. In particular, since the onset of the financial crisis, tax has been a minor consideration in the funding structures of most banks. The evolving regulatory requirements as to the shape and quality of banks balance sheets together with the availability and cost of credit have been far greater considerations and behavioural drivers.

4. Other considerations would also rank as highly as the deductibility of interest costs. These would include: (1) the costs and complexities of the capital raising process; (2) flexibility as to term and other commercial conditions of the funding raised; (3) the effect on the share price of the type of capital raised; and (4) the overall funding profile of the bank.

5. The tax system, therefore, has relatively little influence on the overall funding structures of banks.

1.1.1 How large is this effect?

6. We do not consider that the tax treatment of debt and equity has a significant effect on the amount of leverage undertaken by the banks.

1.1.2 What are the determinants of the size of the effect?

7. As discussed in response to Question 1.1, the determinants over the amount of leverage are primarily driven by commercial and regulatory considerations rather than tax.

1.1.3 Do taxes other than the corporate system introduce any distortion?

8. Notwithstanding our responses above, in addition to UK corporation tax, the bank levy may now play a role in the make-up of a bank’s funding by effectively increasing the cost of leverage. However it is difficult to isolate the deleveraging effect of the UK bank levy from the other dynamic changes in the banking environment such as capital and liquidity changes and profitability challenges. Where the levy remains a centrally managed cost, it is perhaps more likely to have a distortive effect, influencing treasury decisions. Where it is charged out to the business units of a bank, it is more likely to have an effect on the type of business conducted and funding required.

9. To the extent that regulatory resolution reform removes the implicit government guarantee of a bank’s liabilities, the policy basis for a levy on a banks’ balance sheet will diminish.

1.1.4 To what extent is the distortion offset by the personal tax system?

10. Whilst there are differing tax treatments of finance costs of debt and equity under the corporate tax system (although, as discussed in response to Question 1.1, we do not believe that this has a significant effect on the leverage undertaken by banks), the personal tax system will, to some degree, offset this asymmetry by taxing the receipt of interest and dividends differently.

11. However, it should be noted that it is unlikely that a substantial proportion of a bank’s equity and debt would be held directly by UK resident individuals. For example, the equity of a bank will be held by pension funds, asset managers and overseas investors among others. The debt will also be provided by a mix of entities other than individuals such as other banks, corporates, overseas individuals as well as UK resident individuals. Even UK individuals will receive interest in a mix of taxable and non-tax forms (through ISAs).

12. All these different recipients are taxed in different ways and under different regimes so the effect of the UK personal tax position cannot be accurately considered in isolation.

1.1.5 Would regulatory constraints on capital no longer bind if the distortion were removed?

13. Notwithstanding our response to Question 1.1, the removal of a deduction for debt-based financing costs would be unlikely of itself to remove any regulatory constraints on capital. The capital markets still price the risk associated with debt and equity differently and therefore require different levels of return. Achieving the right overall cost of financing in constrained capital markets, and not tax, remains by far the single greatest influence on the make-up of banks’ balance sheets.

1.1.6 Has the responsiveness of the effect to its determinants changed over time?  ("We find that, typically, a one percentage point higher tax rate increases the debt-asset ratio by between 0.17 and 0.28.  Responses are increasing over time, which suggests that debt bias distortions have become more important." de Mooij (2011), The Tax Elasticity of Corporate Debt: A Synthesis of Size and Variations, IMF WP11/95)

14. The relative importance of tax deductibility of debt is inevitably linked to wider factors. Lower UK tax rates inevitably lessen the economic incentive for leverage, as do profitability challenges, but commercial and regulatory considerations ultimately drive a bank’s funding structure.

15. Whilst Ruud de Mooij’s 2011 paper provides an excellent comparison of the economic effects of different systems of taxation, the research was conducted in 2008 against the backdrop of very different economic conditions and very different capital markets conditions. Most importantly, it was not directly addressing the position of an industry whose balance sheets are very heavily regulated.

1.2 What is the best way of removing any such bias? 

16. Our experience is that any bias towards leverage, on the basis of its tax effects, is of such a modest level that it does not warrant a wholesale change in tax system. When compared with any such modest distortion, the benefits of a tax system which is closely aligned to the accounts of the company (i.e. the current system), providing simplicity, transparency and comparability is of far greater value than an attempt to address the minimal leverage distortion.

1.2.1 ACE vs CBIT

17. In summary, in our view, altering the tax system, whether by adopting either an Allowance for Corporate Equity (ACE) or a Comprehensive Business Income Tax (CBIT) system, is not warranted by the limited impact of the tax distortion and would add complexity to the tax system and make the UK less comparable to overseas regimes.

18. A unilateral adoption of either system would lead to a lack of transparency and comparability with other jurisdictions. This, accompanied by the considerable complexities associated with calibrating and scoping either system, would likely outweigh any benefit of removing the modest level of bias that may exist under the current regime.

19. As a general matter, we would observe that changes to the tax system which inevitably lead to a rise in the marginal rate (as ACE would for example, if one of the requirements was to maintain the quantum of UK corporation from the sector), are more likely influence the location not just of capital but of the business activities themselves.

20. For CBIT, the nature of the disallowance for interest expense or potential exclusion arrangements for interest income (again assuming UK corporation tax revenue neutrality from the sector be required) would themselves be extremely complex and may well deter certain financing activities which the UK would prefer to promote rather than discourage (such as overseas lending from the UK).

1.2.2 Are there any other workable options?

21. As noted above, we do not consider that the impact of the distortion is sufficiently significant to justify a change to the system.

22. In theory, a taxation system could apply lower rates of tax to the generation of profits and then differing rates according to the particular use of those profits (higher rates on distribution to shareholders, lower rates (or indeed exemptions) for particular types of reinvestment in the business) is a potential option which would also encourage banks to build their capital base.

23. Any lost revenue on the lower rates on profit generation could be replaced by the tax deducted on profits distributed to shareholders.

24. This would also more directly target the incidence of tax to where it should be borne compared with the current corporate tax system. However, it would also potentially discourage the payment of dividends, and require consideration of potential credit arrangements.

25. This would obviously be a major change to the UK taxation system and would, therefore, require considerable further research.

1.3 Can and should the reform be restricted to banks, as opposed to other financial and non-financial companies?

26. As noted above, we do not consider that the impact of the distortion is sufficiently significant to justify a change to the system.

27. We would note, however, targeting tax reforms on banks alone is becoming an increasingly inexact science as many "banking" activities are now conducted by a broader population of financial service providers, be that consumer finance, wealth management etc. It does, however, remain the case that the core and capital intensive (and therefore most likely to be effected by any such reform) banking activities are provided by a relatively small number of institutions.

28. An ACE based reform could, therefore, to some extent be introduced on a targeted basis (although this would lead to complexity and a lack of clear comparability in the investor market) . A CBIT based reform would be more likely to require broad application, if instances of double taxation are to be avoided.

1.4 What are the administrative and international complexities involved in such a reform?

29. A reform of this scale would require significant modelling and calibrating to ensure it was achieving the desired effect. The deductibility or non-taxation of certain income and funding costs, and the rates at which any deduction or exclusion were set, could result in the reduction of the size and nature of activity undertaken in the UK (as has been seen with the Bank Levy).

30. The interaction with other jurisdictions operating on a different basis would also need to be considered, including any impacts of current double tax treaties.

1.4.1 What would happen if the UK unilaterally introduced an ACE reform?

31. It is not possible to determine or anticipate the effects of an ACE reform without access to the details of the specific proposals. For example, if the overall UK corporation tax rate stayed the same it may result in a movement of commercial activities into the UK.

1.4.2 Are there any international legal obligations preventing a reform in the UK?

32. Again this would depend on the detail of the reform. For example, the reform would need to take account of EU Regulations on state aid to ensure that it could not be read as a government subsidy of UK financial services equity.

1.4.3 How should the balance sheet of UK banks’ foreign subsidiaries, and intra-group debt and equity positions, be treated for the purposes of the reform?

33. Again, this would require a balance to be struck regarding ensuring that intra-group positions do not give rise to unintended consequences, whilst ensuring that EU law was complied with.

1.5 What are the implications of a reform for the Bank Levy?

34. If the Government were to introduce a reform to encourage equity over any form of debt finance, such a reform would alter the taxation burden on banks and hence would require a review of any specific revenue target for the bank levy. It would also likely require the removal of existing exclusions and a recalibration of rates. The impact of such a process on individual banks would also be significantly different.

35. To the extent that the bank levy has any impact on the funding mix of a bank, it rewards equity finance over most forms of debt finance, but treats some forms of debt finance as just as good as equity finance.

36. The exclusions for protected deposits and sovereign repos, and the lower rate for long term liabilities, were built into the design of the bank levy legislation in order to reward bank funding models other than those that rely on short term wholesale market debt funding. But the legislation clearly envisaged that there could be forms of "good" debt funding which are to be encouraged in addition to equity funding.

37. The bank levy is a significant aggregate cost - especially to banks with UK parents and banks which may be in a loss position for corporation tax purposes (meaning they do not benefit currently from corporation tax deductions on interest expense). Even so, in our experience, 0.13% would be unlikely to create a sufficient cost differential to change a decision to fund using equity over debt given the much wider gap in the general cost of equity versus the general cost of debt, absent the levy. Given that the most immediate marginal funding method for any bank is short term debt financing, the more likely behavioural consequences of a bank levy is to discourage balance sheet growth, with an impact on the supply of credit to the wider economy.

1.6 How much would the reform cost?

38. Determining the cost of the reform in terms of the UK corporation tax take would be a detailed and complex exercise, entirely dependent on the precise details of the reform. It would, theoretically, be possible to calibrate it in a way which was UK corporation tax neutral if changes in rate accompanied a change in the base, but that would not be able to accurately factor in resulting changes in behaviour with regard to activity moving in and out of the UK.

1.6.1 How can the distributional effect be neutralised?

39. It would only be possible to understand the distributional effect once the details of the reform were determined.

1.6.2 Is it practical to apply the allowances only to new equity?

40. As above, this would entirely depend on the detail of the reform. If it were a stand-alone notional deduction for equity, applying it only to new equity could well be possible. This would require clear rules as to what types of equity would qualify – for example, it could be linked to further equity issued above a minimum regulatory requirement.

41. If the reform were more complex, with other changes made in order to ensure the overall effect was corporation tax neutral, for example changing the taxation of returns on such equity, or income generated by the funds acquired via such equity this would be considerably more complex. Practically determining which return related to "old" or "new" equity would be extremely difficult if not impossible and therefore may need to be considered on a notional basis.

42. However, as noted below, achieving corporation tax neutrality may be achieved in alternative ways which would be practically simpler, such as an increase in the corporation tax rate, albeit those methods have other practical issues.

1.6.3 Can the rate be increased to offset the narrower base?

43. As noted above, this would be one method of ensuring a theoretical offset. However, increases in rate increase the distortion of the tax system at the margin, including movement in commercial activity to and from the UK. This is difficult to predict, and therefore in practice it would be hard to calibrate the rate needed as changes in rate will change the size of the base in a wider context.

1.6.4 Could the banks issue shares to the Government to compensate it for the loss of tax revenues?

44. Legal certainty, including tax certainty, is an important factor for businesses, when assessing new locations. Hence, given the Government’s objectives to encourage businesses to locate in the UK, maintaining a strong history of a consistent rule of law is extremely important and such a change would be damaging. Whilst share issuance could be used in place of a cash payment or to fulfil a legal obligation to pay, in our view any extension beyond that would be detrimental and should be avoided.

45. In addition, it would give rise to a number of practical issues on an international level, for example such a compensation method may not be a "tax", much like the bank levy, and therefore unlikely to be covered by existing double taxation agreements.

1.7 What are the economic consequences of financial services not being subject to VAT?

46. It is difficult to determine whether the exempt status of VAT leads to more or less net VAT being collected in respect of the activities of financial services businesses. The answer will differ for different financial services businesses depending, in part, on the ratio of business to private customers. Whilst most financial services are exempt from VAT as a matter of EU law, there are some that are subject to VAT (particularly in the investment management area). We focus only on those that are exempt.

47. Notwithstanding the EU’s competence in the field of VAT legislation and addressing the substantive question, there are two key policy reasons for the EU exemption from VAT for financial services. One of these is the difficulty, in many circumstances, of identifying the price charged for financial services in the absence of an express fee charged to customers. For example, it is difficult to identify the consideration for the provision of a 'free' bank account. The second is the desire not to impose a VAT cost on final consumers in respect of the purchase of necessary financial services (although the exemption does also apply to business to business transactions).

48. The economic consequences of exemption compared to taxation may provide a net benefit to HM Revenue & Customs (HMRC) or a net benefit to taxpayers depending on the particular circumstances (we have included, in Annex II, an example to illustrate this point). As financial services are exempt from VAT, this means that the financial institutions are unable to recover VAT on associated costs (such as property, IT, professional advice, etc) used in the provision of exempt services. This VAT cost will, where possible, be built into the prices charged to customers in order that the financial institutions can cover their costs from their revenues.

49. This imposes a net cost on business customers that are able to recover VAT (see Annex II).

50. In contrast, where the customers of financial institutions are unable to recover VAT charged to them (typically because they are final consumers), VAT exemption is likely to be beneficial if the financial institution is able to charge fees in excess of its costs (see Annex II).

51. The reality is more complex as some exempt financial services give rise to a right to recover VAT (e.g. where the customer is outside the EU). In some circumstances, this will lead to a net increase in VAT collected by HMRC and, in others, a net decrease. The situation is further complicated by the added difficulties of identifying the appropriate VAT on all financial transactions.

1.8 Should financial transactions or ‘financial activity’ be taxed?

52. We comment below on the merits, or otherwise, of transaction and activity-based taxes. However, at a more general level, we are not persuaded, in the current environment, that there are strong economic arguments to justify additional taxes that are applied specifically to the financial sector. The UK already taxes certain financial transactions via Stamp Duty Reserve Tax and Insurance Premium Tax. It could be detrimental to impose additional taxes without repealing some of the existing ones. For example, some financial transactions in the UK are already subject to multiple layers of taxation with Stamp Duty Reserve Tax on the consideration for the transaction, UK corporation tax on any profit generated by the transaction and UK Bank Levy on any funding provided to finance the transaction.

53. There is a considerable amount of evidence of the damaging effects which transactions taxes have on turnover in financial markets and therefore core liquidity in those markets. There would, therefore, seem to be limited opportunities to extend financial transaction taxes in the UK without risking the loss of the tax base to overseas jurisdictions, significantly increasing the cost of compliance or risking stability in the financial markets. All these factors were present in the Government’s opposition to the EU’s proposed Financial Transactions Tax. Previous experiments in such taxes, for instance in Sweden, have shown that, unless properly constructed, these types of taxes do not raise revenue and can result in a dislocation in financial markets.

54. As an alternative, Financial Activities Taxes tend to focus upon financial institutions themselves, rather than attach to the transactions that they undertake. We can appreciate that historically certain business models within financial institutions have earned returns beyond normal returns and this has in part been attributable to the existence, implied or actual, of government guarantee to the solvency of these firms. Where such a guarantee exists, we can see the rationale for taxing certain financial institutions more heavily to level the playing field with other sectors. An analogy may be drawn with Petroleum Revenue Tax where there is a specific taxation in relation to the industry due to the inevitable overlap with the state. However, given the Vickers recommendations, we expect this guarantee to be eliminated. Within the ring-fenced retail banks, the contributions made by banks to the Financial Services Compensation Scheme (FSCS) acts as a tax to pay for the guarantee provided to depositors and, therefore, the rationale for additional taxation does not hold.

1.9 What are your views on suggestions that there should be an additional bank levy to bail out future failures, or that certain levels of remuneration be treated in a similar way to distributions?

55. There should not be an additional bank levy to bail out future bank failures. Such levies imply that the taxpayer stands behind banks’ liabilities. "Bail-in" not "bail-out" should be the guiding principle for resolution (as recommended by the FSB) i.e. taxpayers should not bear the cost of bank failures.

56. As regards remuneration, the FSB’s 2009 Principles for Sound Compensation Practices and related implementation Standards are an important step towards achieving global consistency and it is important that further reforms remain consistent with these Principles and Standards to avoid both regulatory arbitrage and unlevel playing fields for banks.

57. One issue that is worthy of greater consideration, at a global level, is the introduction of measures designed to ensure that a bonus is accrued and paid, if, and only if, capital has received a minimum return. Effectively this would transform bonuses from an expense to a form of participation by employees in the profits that the bank has earned above its minimum required return. For example, constraints could be imposed on trading profit-related bonuses, so as to permit payment in cash only when such profits are realised, or at least readily convertible to cash. This could be achieved by permitting payment only out of ‘realised profits’, as defined in the Technical Release issued by the Institute of Chartered Accountants in England and Wales and the Institute of Chartered Accountants of Scotland, TECH 02/10 Guidance on the determination of realised profits and losses in the context of distributions under the Companies Act 2006. Such a proposal would, however, need to be thought through carefully.

2 What are your views on alternative systems to level the playing field?

58. Our views are outlined in the response to Question 1.2.2 above.

3 Do banks’ attitudes to tax planning affect banking standards and culture, and does this have any effect on the wider economy?

59. Banks' approaches to tax planning have become markedly more conservative since the onset of the financial crisis, both with respect to their own corporate position and the remuneration of its employees. The approach taken by any bank towards tax avoidance and the overall management of its tax cost could fairly be said to reflect the overall culture and standards of the organisation which inevitably will evolve over time and sometimes change radically with new leadership. In our experience it is exceedingly rare for a bank's attitude towards tax avoidance to affect its culture or banking standards. For those banks that take a highly sophisticated approach to their tax position, that position is managed under a process with robust controls and approvals, high standards of risk management and disclosed to HMRC in line with all reporting obligations. For the great majority of banks, tax plays a very minor role in the conduct of their day to day banking business when compared to other commercial and regulatory concerns. Indeed most tax departments within banks would list this as one of their greatest challenges both in terms of ensuring tax efficiency of day to day business and achieving business change in line with new legislation and regulation.

4 Do you have any views on the role and purpose of structured capital markets teams in banks? Does the volume and type of structured tax transactions have any effect on bank stability, and did this play a part in the banking crisis?

60. The role and purpose of "structured capital markets" teams varies between institutions and has changed over time, with a number of institutions no longer having such teams.  As a general rule, the role of "structured capital markets" teams is to provide bespoke structuring for specific transactions, not necessarily just tax structuring.  Structuring solely for tax benefits is rare, with the overall purpose of these teams being to increase or ensure the profitability of the transaction in question.

61. The tax consequences of some structured transactions are crucial to their commercial viability, as recognised by specific tax legislation, for example the tax neutral status of securitisation vehicles. However, these transactions are structured from a commercial perspective, with achieving the correct tax treatment being a necessary factor but not the purpose of the transaction.

62. Transactions structured specifically in light of their tax treatment – "structured tax transactions" as opposed to structured products more generally - if undertaken at all, are generally a very small and shrinking part of a bank’s overall business. In our view and at the current time, transactions structured purely for tax purposes do not have an effect on bank stability and, in our experience, were not instrumental in creating the banking crisis.

4.1 Is there any evidence that post-crisis this activity has moved into the sphere of shadow banking and if so, does this have any effect on bank stability and regulation?

63. Some structuring activity has moved into the sphere of shadow banking. In our view, and at this stage, this is not enough to have an effect on bank stability and regulation.  We say this because we have observed a small number of specialist funds having undertaken structured capital markets activity post the crisis.  The volume of these transactions is insufficient, in our view and at this time, to have any effect on bank stability and regulation.

64. However, it is important to recognise that the risk exists and, for this reason, the FSB is looking closely at issues relating to "shadow banking".

5 What are your views on the effectiveness of the Code of Practice on Taxation for banks? Would the Code benefit from having sanctions and if so what should these be?

65. The Code of Practice on Taxation for Banks has had the beneficial effect of providing a catalyst for banks to ensure that their own tax practices and behaviour has good governance and accountability. In that regard it has been effective in providing a consistent basis for HMRC to discuss behaviour across the industry. It has frequently elevated discussions about such issues to more senior levels of management, including the Board. However, much of the work done under the code has been to ratify and endorse that pre-existing behaviours are code compliant. Comparatively little behavioural change has occurred because banks were already substantially operating within the parameters of the code.

66. We do not believe that the code would benefit from sanctions. Primary legislation covering all sectors already has an appropriate range of sanctions to deal with non-compliant behaviour which will be further enhanced by the introduction of the General Anti Abuse Rule, which does contain sanctions.

6 How effective has the Senior Accounting Officer legislation been with particular regard to banking standards and culture?

67. We do not believe the rules were designed with the goal of influencing wider banking standards and culture. They have, however, helped to promote an awareness of tax matters in the wider business and outside the tax department. The senior accounting officer rules were introduced to help ensure that businesses have appropriate tax accounting arrangements, which include personal accountability of the Senior Accounting Officer ("SAO"). This is primarily focused on whether the business has taken reasonable steps to ensure that its end to end systems and processes are sufficiently robust to ensure accurate tax returns and payments. In our experience many banks already took very seriously their obligations for filing returns and making tax payments.

68. Given the significant pressures that banks have faced in recent times with regard to risk control, and the limited scope of the rules, tax accounting arrangements have been a lower priority than some other areas, given the potential exposures are less significant. This could be compared to how such matters are prioritized in other jurisdictions – for example, in Italy similar requirements have higher profile and the relevant legislation makes the Finance Director personally liable for the accuracy of the tax numbers.

6.1 Would the introduction of a power akin to that of the FSA in S166 FSMA 2000 be a useful addition to HMRC’s toolkit?

69. There are circumstances in which such a power may be useful but the effectiveness of this proposal would depend on the envisaged scope. If the S166 FSMA 2000 equivalent is focussed like SAO on the accuracy and governance of the end-to-end systems and processes over those, we would see that as being of limited benefit to HMRC in shaping companies’ behaviours in regard to tax. This is because, in our experience, companies already take their tax compliance issues seriously.

70. The management of tax risks and the governance over them could, however, be improved in some companies by bringing the tax risks more into line with other risks within the business, for example, in tackling customer tax avoidance.

71. In addition, there are a number of areas where a court based legal process is potentially an inefficient and expensive way in which to resolve a dispute – transfer pricing for example. In such instances such a power could be useful.

72. Hence, we believe that the idea of a S166 FSMA power for the HMRC is worth exploring further, depending on the purpose envisaged.

73. It is for HMRC to establish what constitutes taxable income and what constitutes acceptable deductions from that income. HMRC may wish to have recourse to skilled persons to determine whether companies – not just banks - are correctly characterising transactions for tax purposes, but such skilled persons should not be employed to give interpretations of the tax code. In our view, such recourse to skilled persons should be reserved to situations where HMRC has opened an active investigation of a company’s tax return.

74. Safeguards would also need to be put in place regarding how HMRC would use any new powers in investigating tax planning and advice taken by the bank to ensure the proper balance is struck between taxpayer rights and those of the tax authority. In particular, the taxpayer should be informed (similar to S. 167 of FSMA) that matters uncovered by the skilled person may be used by HMRC against the taxpayer in enforcement proceedings.

7 Do we need a special tax regime for banks? If so, what would this look like and what would be priorities for change? Should tax continue to follow accounting with respect to banks? Should the tax system actively seek to influence banking standards and culture?

75. To some extent there already is a special tax regime for banks, as many areas of taxation law apply only to financial institutions or transactions undertaken by financial institutions. Tax alignment to accounting provides simplicity, transparency and comparability and is, therefore, generally encouraged. It is also important that the tax system supports regulatory objectives, therefore, closer alignment would be encouraged rather than utilizing the tax system independently to influence banking standards and culture as it would be unlikely to have the desired effect.

76. As well as a revenue collection method, tax policy is also frequently used as a tool to encourage certain behaviours among tax payers. Well capitalised banks undertaking core-banking activities are critical to the economic success of the country. As a general rule we feel that regulation is a better way to promote these behaviours rather than the tax system, as changes to the tax system to change behaviours can result in unintended consequences.

77. It is, however, important that the tax system supports the regulatory objectives and the current tax rules do not do so on a consistent basis. As discussed above, the bank levy would tend to make equity funding more attractive. However, as outlined above, the tax treatment is unlikely to be the ultimate driver of behaviour when these objectives are not aligned. In our experience, the regulatory requirements and commercial considerations tend to dictate the level of capital held by banks rather than the tax treatment of debt or equity. It would therefore be helpful if the tax system were aligned to incentivise similar behaviour to regulation, so that complying with regulations does not penalize banks from a tax perspective.

8 Are banks exploiting regulatory and information arbitrage between FSA, HMRC and auditors? If so, what is needed to address this?

78. Not as far as we are aware. There are a number of instances where the interaction of the regulatory, tax and accounting treatment of a transaction are crucial to its commercial viability, for example some forms of hybrid capital. However, in our experience, such arrangements are disclosed to HMRC and are transparent, with information being available to the FSA and auditors, and are, therefore, characterised by open information sharing as opposed to information arbitrage. In our experience, we are not aware of circumstances where information arbitrage between the FSA, HMRC and auditors is occurring.

9 Should there be a ‘safe environment’ in which the tax authority, regulator and auditors can share confidential information and concerns, possibly on varying levels of seniority?

79. Generally, a joined up approach to regulation of public interest entities is to be welcomed and, as discussed in response to Question 19, we support greater dialogue between auditors and regulators. However, it will be necessary to understand the remit, scope and safeguards of such a proposal before commenting further: in particular, the common areas of interest.

10 What was the role of accounting standards and reliance on fair value principles in the banking crisis? What does a ‘true and fair view’ really represent to the market?

80. We provide a fuller response to these important questions in Annex 1 to our written evidence.

81. In summary:

i) The adoption by UK banks of IFRS in place of UK GAAP did not have the level of impact that has, on occasion, been asserted. Given the similarities in the relevant literature between UK GAAP and IFRS, the majority of the comments we make about the role of accounting standards and the use of fair values in this and subsequent questions are equally relevant for both sets of accounting standards.

ii) It is clear that accounting standards and the principle of fair value were not the principal cause of the banking crisis. It is also unlikely that the crisis would have been avoided had the accounting been any different. Banks who suffered worst or least did so irrespective of the accounting regime. Financial statements under IFRS, or any other accounting framework, are not designed to stress test assumptions about risks, liquidity and availability of funds, and ensure that a bank has sufficient capital to survive fluctuations in its fortunes. Banks need to have additional procedures to determine if they have enough capital to support the business and withstand unexpected losses that may arise. The higher regulatory capital requirements that will be introduced by Basle III will help to maintain more appropriate levels of capital, but one of the lessons from the crisis is that too much reliance has traditionally been placed on regulatory capital requirements. Nevertheless, various improvements can, and should continue to, be made in the accounting methods used by the banks and other organizations and we are strongly supportive of the efforts made by standard setters, such as the IASB, to do so.

iii) IFRS is a principles-based set of accounting standards, developed following extensive consultation and due process. It requires that financial statements shall ‘present fairly’ the financial position, financial performance and cash flows of an entity, and ‘fair presentation’ under IFRS is regarded as equivalent to a ‘true and fair view’ under UK GAAP. Consequently the application of an IFRS standard will in the vast majority of cases give a true and fair view. Both sets of accounting standards require the application of judgment in selecting appropriate accounting policies, determining impairment provisions and making valuations.

11 What are your views on the current incurred-loss impairment model and its role in the banking crisis? Do you consider that proposals to move to an expected-loss model will address criticisms of the current accounting rules?

82. As explained in our detailed response to Question 10 (in Annex I), the relevant literature demonstrates that the IFRS incurred-loss impairment model is similar to that under UK GAAP and also to that used in most other jurisdictions. It is also not as restrictive as it is sometimes portrayed and its application requires judgment, just as with UK GAAP. It would, therefore, be inappropriate, in our view, to ‘blame’ this model for the crisis. Nevertheless, the current incurred-loss impairment model can be improved, in order to match better the recognition of income and impairment losses on loans, and we strongly support the work carried out by the IASB on developing an ‘expected loss’ model. As originally proposed, this would have required deferred recognition of part of the interest income received on loans, to build up a provision for the impairment losses expected to arise. Based on feedback received, in order to make the model operational, the proposal is now that lifetime expected losses will be recognised on all loans where there has been significant deterioration in credit quality since they were first made, plus a provision will be required for all other loans based on the losses expected to arise in the next twelve months. Based on these proposals, most European banks will, in future, need to make higher provisions.

83. However, although such an improvement will result in the earlier recognition of provisions, an expected loss model will not, in itself, prevent a future crisis and it will inevitably require a greater use of judgment, in the framing of expectations. The crisis in the UK was caused, in part, by assumptions about the development of property values that are now seen to have been optimistic, which not only fuelled the growth of the loan books of a number of banks but would also have informed their expectations of possible loss. Even if the banks had applied an expected loss impairment model, rather than the incurred-loss model, provisions would have had to be considerably increased as the crisis unfolded and expectations were changed over time.

84. For instance, in a study we made in the UK in the summer of 2007, just prior to the crisis, although most respondents built into their impairment calculations forced sale discounts on top of any underlying assumptions about the movement in house prices, six banks assumed that house prices would continue to rise, two that they would remain flat and only one that they would fall [1] .

85. Another example can be seen in the dynamic provisioning model applied in Spain. Although not exactly the same as an expected loss model, the dynamic model seeks to smooth loan losses over the economic cycle by building up higher than normal provisions in good years. Although the resulting provisions were believed to be higher in the years leading up to the crisis than they would have been using an incurred loss model, they were still considerably lower than the provisions made since the crisis with the dramatic and unexpected decline in the fortunes of borrowers and property values.

86. As we have already said in response to the previous question, loan loss provisions should not, and never will, substitute for banks holding sufficient capital to withstand expected and unexpected changes in market conditions.

12 What is the best method of accounting for profits and losses in trading instruments? Are there any alternatives to mark-to-market or mark-to-model that might better represent a ‘true and fair view’?

87. Consistent with the views of both the IASB and FASB, we find it difficult to think of a better alternative to fair value accounting for trading instruments in banks.

88. As explained in our detailed response to Question 10 (in Annex I), there was no significant change in the accounting treatment of trading instruments on the move to IFRS, except for the restriction on ‘day one’ profits. The use of market values to record gains and losses in profit or loss had been established practice in major global banks since the mid 1990s and was recommended by the UK’s banking SORPS on Securities and Derivatives. Without the use of fair values, banks can ‘manage’ their reported results, which will depend upon the timing of realisation of transactions rather than the movements in the market value of trading instruments. What changed during the first decade of the current century was not the accounting treatment but a major increase in the scale of trading in complex derivatives and asset backed securities, that were more difficult to value and where the ability to realise a fair value was dependent on continuing market liquidity.

89. In addition, it should be stressed that many of the fair value losses incurred during the banking crisis, especially on securities positions, arose as market prices plummeted (due in part to the loss of liquidity as investors rushed to exit their positions), rather than a reversal of previously recognised fair value profits.

90. Also, as already mentioned, while there are observers who believe that recognition of unrealised losses on illiquid instruments helped contribute to the loss of market confidence, we believe that they have helped provide transparency and so, in the longer term, will have increased investor confidence.

91. In 2007, the IASB introduced a new financial instrument disclosure standard; IFRS 7. This required disclosure of the valuation methods used and, if fair values are determined in whole or in part using assumptions not supported by prices from observable current market transactions, the effect of using "reasonably alternative assumptions" [2] . The IASB added to these disclosure requirements, in 2009, by requiring banks to classify the financial instruments they hold according to the valuation technique and to provide greater details of how they value so called ‘level 3’ instruments (whose values are not based on observable market data) [3] . It is through these disclosures that users of accounts can understand the extent that banks are exposed to assumptions that cannot be corroborated by liquid market prices. We believe that these requirements could usefully be expanded, by developing guidelines as to how these numbers should be computed in a more consistent manner. One basis for this could be to align the disclosures with regulatory requirements such as the Financial Services Authority’s (FSA’s) recently introduced Prudent Valuation Reporting, which has the aim of quantifying valuation uncertainty using a 90% confidence interval and assessing additional capital requirements based on the level of uncertainty.

92. The effect of more dramatic changes in market prices, particularly in stressed situations, is probably best captured by disclosure of information on banks’ risk positions. Transparency about how an institution views its stress scenarios is useful and helpful to investors and others. Even still, such stress tests would not necessarily model rare events or those considered to happen extremely infrequently as those stress events that causes major losses are rarely anticipated. They are ‘unknown unknowns’.

93. It has also been suggested that unrealised profits should not be recognised on ‘level 3’ instruments, or that they should be recognised in OCI. The latter suggestion would have the merit of including fair value gains and losses on such instruments in the balance sheet, where arguably it is important for banks to reflect asset values as accurately as possible, but not in the main performance statement. One of the challenges with either proposal is that level 3 instruments are rarely traded in isolation and are often economically hedged by instruments that would be classified in levels 1 or 2. In our experience, it would be necessary to expand significantly the scope of hedge accounting, allowing level 1 or 2 instruments to be used to ‘hedge’ level 3 instruments for accounting purposes, so as to avoid recognition of spurious gains and losses, or else to exclude only the element of recorded profit that can be attributed to "unobservable pricing inputs". This would add significant complexity to the accounting process. It is also not clear to us whether such a proposal would significantly reduce the profits recorded to such a degree that it would affect users’ understanding of the risks and financial position of a particular bank, particularly if there is adequate disclosure about valuation uncertainty elsewhere in the financial statements, as discussed above. It is worth noting, however, that unrealised profits on level 3 instruments, in the UK at least, are not regarded as distributable [4] .

94. For all the reasons given above we find it difficult to see a realistic alternative to fair value accounting for trading instruments in banks. The most obvious alternative would be to use the lower of cost or fair value for the entire trading portfolio. Although this sounds more prudent, it would have the disadvantage of distorting the results and financial position of banks by understating profits in good years and overstating them in bad years, while the main determinant of a bank’s reported trading result would be the timing of disposals.

13 Did IFRS accounting standards contribute to a box-ticking culture to the exclusion of promoting transparency and a ‘true and fair view’ of the business?

95. No. The relevant literature demonstrates that IFRS is generally more prescriptive and detailed than UK GAAP. An example would be the rule that an entity may not recognise a ‘day one’ profit when it first enters into a financial instrument, if all the inputs required to value that instrument are not observable in an active market. However, like UK GAAP, IFRS is a principles-based set of accounting standards that require the application of judgement and professional experience by auditors when auditing financial statements prepared by directors. As already mentioned in our detailed response to Question 10, IFRS accounts issued by UK companies are required to give a true and fair view. The move from UK GAAP to IFRS should therefore have had limited impact on how audits of a bank’s financial statements were performed. Auditors had to make similar sorts of judgements about the directors’ choice of accounting policies, financial reporting of loan loss provisions, fair values and distributable profits, whether the financial statements were prepared under UK GAAP or IFRS.

14 Do we need a special accounting regime for banks? If so, what should it look like?

96. The first question that needs to be considered, however, is whether the accounts of a bank should be ‘general purpose’ financial reports, intended primarily for use by investors, lenders and other creditors (as with those of all other listed entities), or whether the primary purpose of banks’ accounts is to achieve other objectives such as to help ensure financial soundness and stability. The IFRS Conceptual Framework notes that parties such as regulators "may find general purpose financial reporting useful, but those reports are not primarily directed towards those other groups" [5] . We believe that regulators can, do and should require additional information to supplement the accounts in order to help regulate the activities of banks and safeguard the interests of depositors. Therefore it is the requirements of investors, lenders and other creditors that primarily need to be met. We will, however, discuss the differing needs of regulators in our response Question16.

97. The historical use of hidden reserves by UK banks, as described in our detailed response to Question 10, was justified at the time on the grounds that banks needed a special accounting treatment. The argument was that it was necessary to hold such reserves because of the significant uncertainties and risks involved in banking and the need to maintain confidence in the banking industry, by avoiding the recognition of large and unexpected losses. However, the practice is not consistent with the objective of general purpose accounts, to provide information to users such as investors on the actual performance of the entity or its true financial position.

98. It is worth noting in this context that the financial instrument IFRSs, such as IAS 39 and IFRS 7, although they have a wider scope than just banks, were developed with banks in mind as major users of the standards. They are more complex standards than would be necessary if designed only for corporate users of financial instruments. Some would argue that there is, therefore, already a special accounting regime for banks.

99. Accounting standard setters have been seeking to develop improved principles-based standards, especially for financial instruments, which respond to the criticisms made following the crisis. These include the work to introduce an ‘expected loss’ loan impairment model. We also welcome the considerable progress made by banks in the quality of their disclosures on the financial instruments they hold and the risks they are exposed to. However, we believe that there is room for continued improvement both in the disclosure requirements of the relevant accounting standards and in the transparency of the disclosures made by banks. The recent recommendations proposed by the Enhanced Disclosure Task Force of the Financial Stability Board [6] , would, in our view, allow meaningful improvement to be made in the quality and consistency of the disclosures.

100. More broadly, better disclosure is not just more disclosure. The financial statements of all companies, not just banks, should tell the story of the business in a clear, readable way. This does not necessarily mean reducing the number of disclosures, but rather focusing on the way the financial statements are presented overall, while still being compliant with relevant accounting standards. We strongly support initiatives such as the European Financial Reporting Advisory Group’s recent Discussion Paper, Towards a Disclosure Framework for the Notes and the work of the FRC’s Financial Reporting Lab, which brings together preparers and users of financial information, to discuss how best to meet the needs of the user community.

15 Are there any interim measures (such as mandatory disclosure) which could be introduced in the meantime?

101. As indicated above, we strongly support efforts to introduce an ‘expected loss’ loan impairment model but we do not believe there is a better method of accounting than fair value for profits and losses on trading instruments. Until the expected loss model is agreed and properly implemented, it would probably be unhelpful to require banks to disclose the potential effects of transition; there will need to be substantial investment in systems to produce the required numbers reliably and, without agreed criteria, significant diversity in application. However, as indicated above, there have already been significant improvements in the disclosures made by banks and we believe that more work should be carried out with a view to further improvements in the quality and transparency of disclosures.

16 How likely are current proposals for improving disclosure and dialogue to be successful (with particular reference to discussion papers issued by FSA/FRC)?

102. Over the last two years, there has been an improvement in banks, the FSA, FRC and audit firms working together to promulgate simultaneous industry-wide implementation of new disclosures. This has made possible a rapid acceleration in industry-wide disclosure improvements, as opposed to former times when changes were cautious and evolutionary. Recent developments in disclosure have generally been designed to improve disclosure and comparability between banks of emerging risks and areas of potential financial loss.

103. There have, however, been stresses and strains. When, in mid 2011, attempts were made to ensure a consistent approach to provisioning against losses on sovereign debt, there were objections in certain international regulatory gatherings, starkly illustrating how difficulties can arise when politics, regulation and financial reporting become interconnected. This was a reminder that prudential objectives may legitimately differ from financial reporting objectives, and that regulators and auditors do not necessarily have the same objectives under their respective legislative frameworks.

104. Since then, the FSA and FRC have issued papers on a variety of topics, mostly in relation to accounting. These have provided a good basis for productive discussion while the papers were still in draft form. We believe the approach has attracted general support from all stakeholders and should be continued.

105. More recently, at a global level, banks, auditors and investors have worked together as part of the Enhanced Disclosure Task Force, organised by the FSB, which issued its recommendations for improving the disclosures by banks in October 2012 [7] .

16.1 Should there be enhanced powers to better align auditors’ incentives with those of the regulator?

106. The primary responsibility of the auditor is to report to the shareholders of the bank on the financial information prepared by the directors in accordance with relevant accounting standards. The reliability of this information can be of considerable use to regulators, and the regulator’s interest in reliability extends from financial statements to regulatory returns.

107. An audit of regulatory returns could provide one form of assurance over their reliability, but there could be additional forms of assurance provided by the bank itself, or indeed gained directly by the regulator’s own investigations. However, the regulator’s primary objective is to ensure prudential safety, and although regulator and auditor frequently travel down the same path, there are times when their paths do and should diverge.

16.2 Should auditors of banks be obliged to have a primary responsibility to the regulator, rather than the client?

108. Notwithstanding the above, if the Companies Acts were to be amended to give an auditor of a UK incorporated bank a primary responsibility to the regulator in respect of its statutory audit, this primary responsibility would, necessarily, have priority over the auditor’s current responsibilities to the shareholders of banks (the ultimate "client"). Whether shareholders of UK incorporated banks should be put in a different position to shareholders of non-UK incorporated banks and other listed companies (particularly as the interests of the regulator and the shareholders may not always be coincident), is clearly an issue for Government to consider, in some detail, with all stakeholders.

109. The question of the primary responsibility of the auditor would also need to be considered in combination with that asked in question 14, as to whether the accounts of a bank are designed primarily to meet the needs of investors, lenders and other creditors, or parties such as regulators. Currently, a bank’s accounts are intended to be ‘general purpose’ financial statements, which are not primarily directed to the needs of regulators. It would make sense to align the primary responsibility of the auditor and the primary purpose of the accounts that he audits.

110. Currently, auditors do have responsibilities to the regulator when they are engaged to perform roles that are directly related to financial regulation rather than statutory audit per se (but could be performed alongside such an audit) - in particular, when an auditor acts as a skilled person under section 166 of FSMA. In certain countries these responsibilities form part of an annual cycle of reporting obligations through the preparation of so-called "Long Form Reports" prepared by the auditor, and provided to the supervisor, and to the board of directors of the bank. However, although there are variations between countries in the detail of the applicable laws, the role of the auditor is to gather and validate information for the use of the supervisor in pursuit of the supervisory objectives.

111. We support the Prudential Conduct Authority’s (PRA’s) and Financial Conduct Authority’s (FCA’s) intention to use section 166 reports as a proactive and targeted supervisory tool (e.g. using section 166 reports to provide the regulator with assurance on specific issues arising from supervisory/desk reviews, including disclosures and aspects of regulatory returns).

16.3 Should regulatory returns be audited?

112. As the ICAEW’s 2010 report [8] illustrates, opinions differ on the usefulness of external assurance on regulatory returns. We continue to believe, however, that the assurance auditors provide on insurers’ prudential returns could, if extended to banks’ returns, provide one form of assurance over their reliability (as noted above, additional forms of assurance could be provided by the bank itself, or indeed gained by the regulator).

113. However, as discussed above, we support a targeted and risk-based approach to assurance on regulatory returns, using section 166 reports (as suggested above), rather than a blanket reinstatement of general assurance requirements in relation to specific regulatory returns. We believe that this approach would chime with the PRA’s judgment-led approach to bank supervision, in that external assurance of regulatory returns could be used as a proactive supervisory tool – the timing and parameters of which would be scoped by the PRA – to supplement supervisors’ own assessments – either to follow-up specific concerns or for quality control purposes - rather than a standard reporting requirement.

17 Should the scope of audit be widened so that auditors can better express a broader view of the business? For example should auditors comment specifically on issues such as remuneration policy, valuation models or risk?

114. If stakeholders believe that auditors should express, publicly, a "broader view of businesses", then the scope of a statutory audit, as set out in the Companies Acts would need to be widened accordingly. However, if the enhanced reporting envisaged (for example on issues such as remuneration policies, valuation models and risk) is intended for regulators, then auditors could give private reports to regulators currently, without a change in the framework for statutory audits (although, depending on the opinions requested, a significant amount of work might be needed to develop criteria and guidelines for auditors to follow).

115. Hence, there are three elements in our answer to this question which need to be distinguished:

1. Enhancing the value of the current auditing model

2. Expanding the scope of the audit; and

3. Leveraging the insight and knowledge of the auditor by prudential supervisors

Enhancing the value of the current auditing model

116. The primary product of an audit is information, upon which the auditor has provided assurance, being made available to third parties about the financial statements as a whole. Except in very rare cases where legal questions arise, or in the rare occurrence of qualified audit opinions, the auditor does not pick out specific areas of the financial statements for separate comment in the published audit report (although auditors already opine, indirectly, on the valuation models used and disclosed risk positions, as part of their audit reports).

117. That said, we strongly support meaningful change to the audit report to increase its usefulness and the value of information in it. The International Auditing and Assurance Standards Board (IAASB) is currently reviewing the audit report with a view to making improvements in this regard. We support many of the changes proposed such as the concept of introducing auditor commentary into the report. This would require auditors to highlight matters that, in the auditor’s judgment, are likely to be most important to users’ understanding of the financial statements and draw attention to management’s disclosure of those matters. As well as pointing to the disclosures, auditors also need to provide an indication of why they believe the matter is important to users’ understanding. For further details of our point of view please see our November 2012 Point of View on Auditor Reporting [9] .

118. The FRC is also active in this area. It is reviewing the auditor’s report to identify how its value can be increased including disclosing information about matters such as audit scope and materiality. It will shortly publish next steps in the Sharman Review on going concern, which we expect to cover the auditor’s responsibilities in this area. The FRC’s upcoming review of its guidance for directors on internal controls including risk may also be relevant. We await further details of this work with interest.

119. It is also both normal, and an expected outcome of the audit process, that the auditor will discuss in detail with the audit committee, a wide range of issues and judgments that have gone into the preparation of the financial statements. The 2012 ICAEW report "Enhancing the Dialogue between Bank Auditors and Audit Committees" [10] sets out these points in greater detail. The areas suggested in the question for additional comment (remuneration policy, valuation models, risk) might well be the subject of discussion at the bank’s audit committee, and the Commission will be aware of 2012 changes to the UK Corporate Governance Code, supporting guidance and UK auditing standards designed to provide shareholders with more information about the subjects discussed at meetings of a company’s audit committee. The intent behind such public disclosures by audit committees would be to help shareholders gain a better understanding of the judgments made by a company, and of the scrutiny given to such judgments by members of the audit committee and the auditors. We are fully supportive of these developments.

Expanding the scope of the audit

120. Unlike developments to enhance the value of the current audit model, consideration of how the scope of the audit (and the underlying corporate reporting) might be expanded beyond its current form are less well advanced. Projects like the ICAEW’s Audit Futures [11] initiative are to be welcomed. Its strategy is to innovate, promote the value of audit and assurance to all sectors and address the criticisms leveled at the audit profession and the current audit model. It does this by convening different stakeholders to create opportunities for dialogue and for collaborative solutions to emerge. It is still early days but we anticipate it will cover the question of whether and how the auditor could give a broader view about a company’s business publically. We are active participants in the initiative and are pleased to be one of its funders.

121. When looking at an extension to the scope of audit there are two important principles that need to be borne in mind. First, where an opinion from the auditor is sought it should be on information provided by the company rather than by the auditor unilaterally. It is an important governance principle that companies are run by the directors (not auditors) for the benefit of shareholders. Therefore it is the directors who should decide what is disclosed. Accordingly, in developing the auditor’s role, corporate reporting would likely also have to develop.

122. Secondly, some information is more capable of being assured than other information. For example, it would be more straightforward for an auditor to give an opinion about the adequacy of a company’s systems and controls for risk management and disclosure of key risks rather than to give an opinion about whether the key risks disclosed are the "right" risks and/or whether the mitigation steps taken are adequate.

Leveraging the insight and knowledge of the auditor by prudential supervisors

123. We fully accept that, although the primary role of an auditor is to substantiate information communicated by others, the auditor is frequently able, based on knowledge and other experience, to offer private insights about matters such as remuneration policy, valuation models or risk that can be of value to management, boards of directors, regulators and other stakeholders.

124. Transforming these communications into verified public information would be a wholly different exercise. For example, auditors have discussions with supervisors [12] on a variety of issues, some arising from specific instructions to carry out investigative work on behalf of a supervisor, and others based on information gathered in the ordinary course of other work, such as the audit. These exchanges occur frequently between supervisor and auditor, but are not made public, which makes possible a very different (and valuable) type of exchange from what would be possible by public disclosures.

18 What would be the effect of using return on assets as a performance measure in banks, as opposed to return on equity?

125. A focus on ROE by institutional investors and banks has, in our experience, almost certainly incentivized banks’ management to pursue strategies that have significantly increased the leverage of financial institutions, heightened the volatility of their returns and incentivised short-termism in attempts to boost equity returns.

126. It is important, therefore, that banks and therefore also investors shift focus to risk adjusted metrics. This will be easier for banks, which can look at a variety of metrics including return on economic capital and return on risk weighted assets.

127. It is also important to consider the return after allowing for potential for losses under stress scenarios. Institutional investors too need to ask more questions about the risks being run and stress testing carried out to assess risk adjusted returns. Return on assets is not, however, a risk adjusted measure.

19 Are the amendments to the Financial Services and Markets Act 2000 regarding dialogue between regulator and auditor sufficient, or does further work need to be done in this area?

128. In our opinion, the legislative and supervisory framework now being put in place in the UK makes clear the roles and responsibilities of all parties, and we do not believe there is a missing element that needs to be created by additional legislation. However, there are opportunities to "fine tune" the legislation to seek to remove real or perceived obstacles to communication.

129. It is generally agreed that the system under the former Bank of England regime worked better than the system of recent years in terms of communications among banks and those involved in supervising and auditing them. The Bank of England did so with a great deal less supporting legislation. To be fair, though, it is not possible to say how the financial crisis might have worked out if the previous regulatory framework had still been in force: perhaps other weaknesses and flaws might have emerged. However, it seems to us that among the various stakeholder groups involved in banking supervision, every effort is now being devoted to establishing and maintaining a good system of mutual understanding. The new system of meetings and communications between banks, supervisors and auditors should be left to continue to strengthen. Open and clear channels of communication are vital.

130. In 2010, the audit profession was instrumental in identifying concerns with respect to communications between regulators and auditors. We were represented in the working group that, following the ICAEW’s proposals, [13] contributed to the development of the FSA’s May 2011 "Code of Practice for the relationship between the external auditor and the supervisor" (the Code). We fully support the Code and believe that its implementation was an important step in the development of increased levels of cooperation between auditors and prudential supervisors. Hence, whilst the underlying detail is not yet available, we welcome the Government’s amendments to the Financial Services Bill, in the House of Lords Report Stage, [14] that will, effectively, "hardwire" the overarching principles of the Code into primary legislation.

131. To be fully effective, though, the dialogue between auditors and prudential supervisors must be a two-way process. Currently, Principle 3 of the Code contains a "presumption" that the supervisors will share information with auditors in the interest of contributing to higher quality audits - subject to any "restrictions on the information the supervisor can share with auditors and the circumstances in which it can be shared". We support, therefore, the introduction of new section 339A of FSMA, which will place the PRA under a statutory duty to engage with auditors of PRA-authorised persons and require it to issue and maintain Code of Practice (a copy of which must, inter alia, be provided to Treasury and laid before Parliament) describing how it will comply with its duty. We believe that this statutory duty is a key development in enhancing the effectiveness of the dialogue between auditors and prudential supervisors.

132. As we noted in responding to the joint FCA/FSA June 2010 Discussion Paper: Enhancing the auditor’s contribution to prudential regulation (DP10/3), one of the reasons given historically by the FSA for not initiating a conversation or notifying auditors of particular concerns is that there are legal constraints which prevent it from so doing.

133. Section 348 of FSMA contains restrictions on the disclosure of confidential information; however, section 349 provides certain exemptions, including disclosures made under the Financial Services Act 2000 (Disclosure of Confidential Information) Regulations 2001. We note that one of the permitted disclosers, in Schedule 1 of the Regulations, is to "an auditor exercising functions conferred by or under the Act" where the information relates to "those functions" and "an auditor of an authorised person appointed under or as a result of an enactment (other than the Act)" where the information relates to "his functions as such".

134. In responding to the FSA’s March 2011 guidance consultation on the draft Code, EY sought to "encourage the FSA (or its successor bodies) to review in due course how the Code works in practice, to satisfy itself (themselves) that as much relevant information as possible flows in both directions." Particularly given that the duty in new section 339A(1)(a) of FSMA will apply only to the sharing of information "that the PRA is not prevented from disclosing", we continue to believe that a review would be helpful. We believe that this review should seek to ensure that the need for the information to be "relevant to the fulfilment of the auditor’s statutory duties" does not inappropriately limit the information prudential regulators are able to disclose.

135. We also continue to believe that section 348 of FSMA, The Disclosure of Confidential Information Regulations and the FSA’s memorandum on the issue, [15] should be reviewed both with a view to producing additional guidance for supervisors or seeking amendments to the legislation. We would be happy to contribute to such an exercise.

8 January 2013

ANNEX I

Detailed response to Q10

Introduction

136. First, by way of context, the accounting by UK banks has evolved considerably over the last 40 years. The UK GAAP applied by the UK banks in the decade or so prior to the adoption of IFRS was very different from that used by the banks prior to the late 1970s. Traditionally, UK banks were permitted to maintain 'hidden reserves', to smooth reported profits and provide a capital cushion to absorb unexpected losses. The clearing banks were forbidden to use hidden reserves in 1970 but for much of the next decade applied a formula to smooth the recognition of loan impairment costs and the gains and losses on the sale of securities. Meanwhile, so-called 'exempt banks' continued to maintain hidden reserves through the 1980s. While hidden reserves and smoothing formulas may have helped to maintain the stability of the banking system, the accounts did not purport to present the banks' true financial position.

137. Second, it has been suggested by some that the change from UK GAAP to IFRS for listed companies, in 2005, has resulted in much less prudent and less appropriate accounting, particularly for financial instruments. However, the treatment of financial instruments under IFRS is not as different from that under UK GAAP as has, on occasion, been asserted. There was no UK accounting standard dealing with these topics, but there were Statements of Recommended Practice (‘SORPs’), developed for banks by the British Bankers’ Association ("BBA") and ‘franked’ by the Accounting Standards Board in the 1990s [1] . For banks these SORPs constituted UK GAAP. The BBA SORPs set out principles for determining provisions for loan impairment and the use of fair values for trading securities and derivatives that were very similar to those introduced subsequently by International Accounting Standard 39 Financial Instruments: Recognition and Measurement (‘IAS 39’).

138. For example, the relevant literature indicates that the UK GAAP loan impairment model for banks was also based on ‘incurred’ rather than ‘expected’ losses, similar to that in IAS 39. Although the SORP on Advances used the concept of a ‘general provision’, this was made "in respect of impaired advances which have not yet been specifically identified, but which are known from experience to be present" [2] . The SORP "emphasised that the general provision relates to impairment already existing in the advances portfolio at the balance sheet date. It does not relate to advances which at the balance sheet date are subject to no more than normal credit risk, but which in the nature of things may become impaired in the future" [3] . As, described in more detail below, general provisions permitted under UK GAAP were replaced by ‘collective provisions’ made under IAS 39.

139. There are differences in the precise wording of the loan impairment models under UK GAAP and IFRS, so that the published accounts of UK banks recorded changes in their levels of recorded provisions on transition to IFRS in 2005. However, the changes went in both directions; while some banks such as HSBC and HBOS recorded a reduction in provisions on transition, Barclays, RBS and Lloyds all recorded an increase in their loan loss provisions [4] .

140. Meanwhile, securities held for trading were recorded at fair value through profit or loss under UK GAAP as well as under IAS 39; the SORP on Securities had required all securities that were not held for longer term ‘investment’ purposes to be marked to market. Also the Derivatives SORP already required all derivatives to be recorded at fair value, except those used for hedge accounting purposes. In some respects, IFRS is more prudent than UK GAAP, an example being that it does not allow banks to recognize a revaluation ‘day one’ profit when first entering into a transaction, unless the fair value of the instrument is calculated only by reference to observable market trades.

141. The principal area in which IFRS increased the use of fair values is the ‘available for sale’ category of securities traded in an active market that, while not being held for trading, are also not expected to be held to maturity. These are required to be recorded at fair value but any gains and losses are recorded through ‘other comprehensive income’ (‘OCI’), in effect a revaluation reserve. If the security is sold, or considered impaired, any gains or losses are recycled to profit or loss from OCI. The effect of this category in the banking crisis was to highlight that many debt securities that (under UK GAAP) would have been recorded at amortised cost, less any impairment provisions, were (under IFRS) reported at much lower market values and (if regarded as impaired) with larger losses recognised in profit or loss. Prior to the crisis, any unrealised gains on ‘available for sale’ instruments would have been recognised in OCI, did not contribute to banks’ regulatory capital and were not distributable.

142. Given the similarities in the relevant literature between UK GAAP and IFRS, the majority of the comments we make about the role of accounting standards and the use of fair values in this and subsequent questions are equally relevant for both sets of accounting standards.

143. Third, the loan impairment model used in the BBA SORP and under IAS 39, while described as an incurred loss model, does not restrict provisions to those loans that have already failed. The impairment criteria set out in IAS 39 not only include loans that are in default but also loans in respect of which there is objective evidence of loss events which have an impact on the estimated future cash flows. It therefore requires that banks make a ‘collective’ provision for losses on those loans which, while they are not yet individually considered impaired, are part of a group of loans for which there is "observable data indicating that there is a measurable decrease in the estimated future cash flows…since their initial recognition" [5] . The guidance to IAS 39 also requires that "the process for estimating impairment considers all credit exposures, not only those of low credit quality [6] ". The collective assessment under IFRS is similar to what was termed a ‘general provision’ under UK GAAP [7] .

144. Fourth, there is much more required disclosure under IFRS than pre-2005 UK GAAP - which has tended to make financial statements generally more transparent.

The role of accounting standards and the principle of fair value in the banking crisis

145. It is clear that accounting standards and the principle of fair value were not the principal cause of the banking crisis. It is also unlikely that the crisis would have been avoided had the accounting been any different. The causes of the crisis can now be seen to be poor judgments by banks of the risks that they were taking on, whether due to the trading positions they entered into, the loans that they made or the methods they used to finance their activities, relative to the amount of capital they held. Banks who suffered worst or least did so irrespective of the accounting regime; those banks in Canada and Australia, for instance, who fared relatively well, applied similar accounting principles to those in Europe and the USA who did not.

146. Financial statements under IFRS, or any other accounting framework, are not designed to stress test assumptions about risks, liquidity and availability of funds, and ensure that a bank has sufficient capital to survive fluctuations in its fortunes. Banks need to have additional procedures to determine if they have enough capital to support the business and withstand unexpected losses that may arise. The higher regulatory capital requirements that will be introduced by Basle III will help to maintain more appropriate levels of capital, but one of the lessons from the crisis is that too much reliance has traditionally been placed on regulatory capital requirements. Nevertheless, various improvements can, and should continue to, be made in the accounting methods used by the banks and other organizations and we are strongly supportive of the efforts made by standard setters, such as the IASB, to do so. Disclosures about such requirements, as are currently given under ‘Pillar 3’ of Basel II, either in the financial statements or on banks’ websites, will also better inform users of the accounts.

147. IFRS has been criticised for requiring losses to be recorded based on market values that were viewed by some as short term and not necessarily connected to ‘fundamental’ values. Critics go on to argue that the recognition of such losses has, consequentially, magnified the effects of the crisis and helped lead to the erosion of banks’ regulatory capital. However, the relevant literature demonstrates that similar losses would have, for the most part, been recognised under UK GAAP and it is instructive that banking regulators did not choose to reverse such provisions for the purposes of calculating regulatory capital. Also, recognition of these losses helped to ensure the transparency of financial reporting and so, we believe, increase investor confidence in the longer term.

148. Nevertheless, various improvements can, and should continue to, be made in the accounting methods used by the banks and other organizations and we are strongly supportive of the efforts made by standard setter, such as the IASB, over the last four years, to strengthen accounting standards. We make further comments about the incurred loss impairment model and the use of fair value in response to Questions 11 and 12.

True and fair view

149. Like UK GAAP, IFRS is a principles-based set of accounting standards that requires the application of judgment and professional experience by the directors when preparing the financial statements, and by auditors when performing their audit. Since IFRS has been applied, a body of supporting literature has been developed which seeks to ensure consistency of application globally and prevent abuse.

150. The Conceptual Framework to IFRS sets out that if "financial information is to be useful, it must be relevant and faithfully represent what it purports to represent. The usefulness of financial information is enhanced if it is comparable, verifiable, timely and understandable." It goes on to say that faithful representation has three characteristics: financial information needs to be complete, neutral and free from error. IAS 1 Presentation of Financial Statements, meanwhile, requires that "Financial statements shall present fairly the financial position, financial performance and cash flows of an entity. [1] "

151. There is no specific mention of a ‘true and fair view’ in IFRS. Nevertheless, it has always been recognised in the UK that financial statements prepared under IFRS should give a true and fair view, as was emphasised by the UK Financial Reporting Council (FRC) in its document on the topic in May 2011. This refers to the opinion of legal counsel that ‘fair presentation’ under IFRS is equivalent to a ‘true and fair view’ under UK GAAP [2] .

152. The FRC document also states that application of an IFRS standard will in the "vast majority of cases" give a true and fair view, since the development of the standards has followed extensive consultation and full due process [3] . This involves input from investors, regulators and auditors as well as preparers, with the consequence that there are checks and balances which guard against undue influence being brought to bear by any one constituent. As a result, the standards generally represent the best view of standard setters and their constituents, at a point in time, of how to account for particular types of transactions. Under UK GAAP, on most occasions when an entity applied ‘a true and fair override’ it was as a result of a conflict between the required accounting practice and Company Law, rather than the perceived need to depart from an accounting standard. However, the FRC indicates that it may be necessary to depart from an IFRS standard in rare circumstances (as also set out in paragraph 19 of IAS 1). According to the FRC, "These circumstances are more likely to arise where the precise circumstances are not covered by a relevant standard [4] ." The document also highlights the continuing importance of the application of judgment, such as in selecting appropriate accounting policies and making valuations.

153. There is also no explicit reference in the IFRS Conceptual Framework to ‘prudence’ as a fundamental accounting concept. However, as stated by the FRC, "The concept of prudence continues to underlie the preparation of accounts... [5] " The FRC’s document explains how the concept of prudence in UK GAAP has evolved to that of ‘neutrality’, primarily to reflect a growing concern about profit smoothing through the recognition of prudent, but not yet warranted provisions (for example costs such as those associated with future reorganizations). The FRC adds "Whilst there are examples where UK GAAP gives rise to lower equity than IFRS the reverse is also true" [6] .

ANNEX II: Question 1.7 - VAT example

154. Suppose a manufacturer borrowed £1m from a bank for a two year period with interest totalling £100,000 (if VAT is ignored). If, unlike the current position, the provision of the loan was subject to VAT, the bank would charge interest of £100,000 + VAT of £20,000 and the bank would recover all of the VAT that was properly allocated to the making of the loan. The manufacturer would recover the VAT of £20,000 in full and there would therefore be no net VAT cost to either party associated with the loan. Under the current model of exemption, the bank would seek to charge interest of £110,000 (assuming it had incurred VAT of £10,000 on costs associated with making the loan). Since this extra £10,000 would not be VAT, the manufacturer would be unable to recover it. HMRC would be £10,000 better off (since it did not have to refund to the bank the VAT of £10,000 that it had incurred) and the manufacturer would be £10,000 worse off as the net cost of interest would be greater than had it been subject to VAT.

155. Under the above example but with a final consumer as the borrower, the effect of exemption would be to deny the bank the recovery of £10,000 of VAT but to save the customer £20,000 of VAT. This would leave HMRC worse off by £10,000 compared to if the interest was subject to VAT and give the customer a net benefit of £10,000 compared to the position if VAT was charged.