Professor Eamonn Walsh:

I thank the Chairman for his introduction and the committee for the opportunity to assist the inquiry. I was asked to consider the role of accounting in bank crises and also external auditors. I have prepared some brief remarks on these two topics which, in the interests of brevity, I have distilled from my published statement. I would be happy to elaborate on any of the points I have raised. I am also happy to discuss other accounting issues of relevance to the banking crisis.

When we think about banking crises, accounting is largely a silent bystander. A bank crisis generally involves rapid loan growth alongside which we get concentrations of risk. Often this rapid expansion is funded from volatile sources. When one is in this expansionary mode, if a regulator or anybody else questions the business model, management will dismiss them because they will be able to point to excellent accounting profits and, as a result, great contributions to capital. This pattern of behaviour was evident prior to 2008 in Ireland. Our banks were among the most profitable in Europe. It was widely believed that in what appeared to be challenging scenarios there were sufficient cushions to withstand risks. Balance sheets and income statements formed the basis for this analysis so accounting was clearly implicated in the misperception of risk. By 2009 it was apparent that the capital cushions available to Irish banks were entirely inadequate. Some financial institutions had liabilities that exceeded their assets; they were insolvent.

Balance sheets and income statements are prepared using accounting rules or accounting standards. In general these are known as international financial reporting standards. They are a large body of rules which indicate how one should measure assets, liabilities and income for an enterprise. However, during banking expansions these accounting standards may be a little unhelpful and they will serve to over-estimate the profits that are reported to external investors and lenders. If we start thinking about bank lending which is going to be the principal asset for banks, and the Irish banks in particular, essentially the big problem is estimating loan impairments. Loan impairments are the amount of the loans that will not be repaid and it is a challenging accounting problem. Impairment may be understood as something not performing as anticipated. If we think about it in American terms, one would speak about someone being visually impaired - it means the eyes are not working as originally anticipated. The notion of impairment means that things are not working as originally anticipated.

With bank lending, if one knew exactly how much the bank was going to lose the accounting would be straightforward. For example, if a bank engaged in very risky lending and we knew in advance just how risky that was and the risks entailed, impairment accounting would be very easy, because we would say we expect to lose all the money we lent to a particular customer, therefore, we should buck a loss and decrease profits and decrease capital. The problem is that this scenario is highly unlikely. If we lent money to a customer and we knew we were going to lose it all immediately we would not lend in the first place. Lending is based upon the proposition that we do not expect to lose our money when we lend money to a particular identified customer. The problem then is that we must engage in some estimates of the defaults that might occur. Given that there are estimates and judgment calls involved in determining these impairments, it is necessary to have additional guidance. The accounting standards as they stand require an entity to assess at the end of each reporting period whether there is any objective evidence that a financial asset or a group of financial assets is impaired. In other words, one is required if one is preparing financial statements for a bank to sit down and ask at the end of an accounting period whether there is objective evidence that an impairment has occurred. The rules go on to state that a financial asset is impaired and impairment losses are incurred if, and only if, there is objective evidence of impairment as a result of one or more events that occurred after the initial recognition of the asset. In other words, one must point to some event that has occurred since we originated the loan that constitutes objective evidence that we have made a loss.

If we believe that a loss has been made it is necessary to estimate the impact of that loss and one is required to produce a reliable estimate of the amount of that loss. The most counter-intuitive part of all of this is that the standard explicitly states that losses expected as a result of future events no matter how likely are not recognised. In other words, if the dogs on the street expect that losses will occur they are ignored by accounting standards. One is required to have objective evidence that the loss has been incurred rather than a belief that events will occur in the future which will endanger these loans. Therefore, it is a very conservative definition of impairment since it requires objective evidence that a loss has been incurred. As a result, impairment accounting is pro-cyclical. What this means is that in a period of expansion as a bank expands its loan book, the bank will appear far more profitable because as it originates the loans it does not expect to lose money, it has not incurred a loss and so the bank will appear incredibly profitable. In turn that means it will have additional capital so it will appear that it has additional cushions against future losses. In a period of contraction the reverse happens.

When the bank expanded it made lots of loans. When there is contraction it becomes apparent that some of these loans are impaired. That immediately means a swing from having excessive loss of profits to excessive losses. That is what we mean by procyclicality.

There are two additional features of accounting policies which lead to even greater procyclicality, one of which is the accrual of interest.

It means that if one lends money to a customer and agrees that they do not need to repay interest immediately then instead the interest will be rolled up on the loan. That roll-up of interest will be recognised as a profit even though the cash has not been received from the customer.

The second thing that can juice up the amount of income that is reported during the expansionary phase is the existence of loan arrangement fees - in other words, some agreement whereby the lender will give money to the bank, at the origination of the loan, as compensation for the origination costs of that loan. It does mean that accounting rules give rise to excessive reporting of profits when we are expanding credit and it will also lead to substantial declines later in the cycle. This cycle was quite evident in Ireland. We had exuberant new lending which gave way to no lending at all because we went from having very high profits to almost no profits at all. Nevertheless, it is difficult to conceive an outcome that would have been very different. Even had accounting rules permitted the use of expected losses, as distinct from incurred losses or other alternatives, I believe that profits would have exhibited a similar pattern.

Impairment accounting is largely diversionary for four reasons. First, the rules are well understood by informed users of bank financial statements. In other words, what I have said today is not new and has been known for decades.

Second, individual banks could create additional provisions despite the accounting rules. There is nothing to prevent a bank from saying "We believe, to give a true and fair view of our results, that we should book additional provisions this year."

Third, there is nothing to prevent the Financial Regulator from insisting upon additional provisions or enhanced disclosure. The regulator did exercise such powers.

Fourth, the expected loan losses during the growth phase in Ireland were likely to have been low anyhow. No matter how one tries to cut it, during the period 2004 to 2006, inclusive, it is very difficult to conceive of any type of a mechanism that would have resulted in a large decrease in the reported profits of banks.

While these measurement rules are quite different it is important to realise that banks also make disclosures. This is a key part of financial reporting, and it is not just what is in an income statement and a balance sheet, but also the additional disclosures in the financial statements. Additional disclosures help one to understand exactly what is going on in terms of reported income. In my opinion, the disclosures concerning the increased risks within Irish banks were inadequate during the period.

Through the lens of 2002 balance sheets - which were balance sheets dominated by residential lending - users could have easily concluded that increased profitability was synonymous with increased cushions but a seismic shift occurred during 2002 to 2007. There was a big change in the composition of loan portfolios for Irish banks. Rather than residential mortgages dominating property landing, commercial and development lending became almost 50% of property lending. In other words, the composition of the assets changed very dramatically. There were concentrations of risky lending that displaced less risky home mortgages. The distribution of the shift is also significant. It is clear that these additional risks were not distributed equally across banks in Ireland at the time. Further, a significant proportion of the loans were concentrated among a relatively small number of borrowers.

In summary, I believe that the real challenge for external users of financial statements was understanding the changing nature of the risks on balance sheets before 2007. Had users fully understood the increased exposure to commercial and speculative lending, and that it was concentrated amongst a very small group of borrowers, it would have been far easier to realise that the quality of profits had declined and that capital mattresses, rather than cushions, were required.

The second point I was asked to talk about was bank audits. What do auditors do? Directors are responsible for preparing and approving the financial statements of a company or a bank and then auditors issue an opinion on the financial statements. For example, an audit report will state something like the group's financial statements. It will give a true and fair view, in accordance with international financial reporting standards as adopted by the European Union, of the state of the group's affairs as at 31 December 2015 and of its profits for the year then ended. The first thing auditors do is express an opinion that the financial statements have been produced in conformity with international financial reporting standards.

Bank auditors have some additional responsibilities to a regulator and are required to bring certain matters to the attention of a regulator. That means there are some additional responsibilities between an auditor and a financial regulator.

In addition to these responsibilities, a regulator has the power to request information from an auditor. There is a particular additional set of duties with respect to bank auditors. For example, bank auditors would have been required to perhaps produce reports of any breaches of prudential sector lending limit guidelines. As Nyberg concludes, in the majority of cases the auditors did not report regulatory sector lending limit excesses to the Financial Regulator. Even if all excesses had been reported, it appears unlikely that any action would have been taken by the Financial Regulator who was already aware of, and not concerned about, such excesses. In other words, there were alternative reporting mechanisms between the banks and the Financial Regulator that are quite independent of annual financial reports prepared for shareholders and lenders. While I have not had access to the auditors' communication with the Financial Regulator, Nyberg states that the auditors clearly fulfilled this narrow function according to existing rules and regulations.

The committee might wish to know what else auditors do apart from expressing this opinion. For say loans and receivables, there would be a general expectation that auditors should go along and look at a sample of some of the lending files within a bank. As part of their audit they would do the following: ensure that lending policies are adhered to; review concentration reports and related party loan reports; establish evidence of any collateral assigned to the institution; check the financial condition of co-signatories and guarantors, examine past experience with the enforcement of guarantees, and confirm terms with the guarantor; compare loan amounts with appraisals; and ensure that construction loans are correctly classified as loans rather than real estate investments. For significant construction projects, they would ensure that advances are based on progress and that there are offsite-onsite inspections to verify the collateral and the progress with the construction.

Auditors should also assess management's loan reviews for impairments. Audit procedures should establish that management has looked beyond the collateral to identify potential borrower weaknesses, to identify if collateral appraisals are adequate and that up to date borrower financial information is available.

Auditors are also expected to review performance against the original loan agreement and be alert to any biases, for example, loans to public figures or personalities.

In summary, when published financial statements are prepared in accordance with international financial reporting standards, auditors also play a valuable role in assuring the veracity of the loans and receivables on a bank's balance sheet.

I shall make some concluding observations. Between 2002 and 2007 the aggregate loans in Ireland grew rapidly and exhibited significant concentrations of commercial lending, and speculative commercial lending. However, these concentrations varied across banks and there were also significant individual borrower exposures which gave rise to exponentially greater risks. Contemporaneous knowledge of these loan portfolio risks would have alerted external users to the sources of bank profits and their sustainability.

While there was no requirement to report these risks in the published financial statements, especially for financial statements prior to 2007, there were requirements to report capital adequacy, liquidity, impairment, large exposures and sectoral limits to the Financial Regulator. It is an empirical matter as to whether these amounts were correctly reported to the regulator and to the boards of financial institutions.

I thank the committee members for their attention and I would welcome any comments or questions you might have.