Address by Governor Patrick Honohan to Institute of Certified Public Accountants in Ireland

27 October 2010 Speech

Current and recent discussion on Ireland’s macroeconomic and financial situation has focused to a remarkable extent on accounting numbers and so it is timely that I should have this opportunity to deliver a lecture to the Certified Public Accountants in Ireland.

As accountants, you will want any accounts to show a true and fair view of the economic entity for which the accounts are being drawn up. I would like to discuss some of the ways in which the normally used accounts have not helped us as much as they might in understanding and measuring the scale of our crisis in Ireland.

Let me start with the government finances. After the dramatic budgetary correction of the late 1980s, this was not expected to be a problem. Indeed, as shown in Chart 1, Ireland ran a balanced budget on average for two decades before the collapse of 2008. It satisfied the deficit criterion of the Stability and Growth Pact (SGP) of the European Economic and Monetary Union before it was invented. And there was no drift: indeed the deficits became smaller and turned into surpluses in the early years of the new millennium.

The Government debt to GDP ratio fell throughout this period reaching a low point of about 25 per cent in 2006-7. Interestingly, in most years the total value of debt outstanding actually increased, though with growth in GDP, both real GDP and the price level component, the ratio to GDP fell steadily.

So what happened in 2008? There was no really significant policy change – OK there had been some late-cycle relaxation in budgetary discipline in 2006-7. But mainly it was that the policies that had delivered steady surpluses suddenly stopped delivering. Obviously we know what happened to the economy: the property prices had stopped rising at the end of 2006, capital gains were down, transactions had slowed dramatically, construction had contracted. But my point is that, as an indicator of the sustainability into the future of the public finances, the SGP rules were not adequate. A country could run a surplus for years and yet the structure of its tax receipts and spending rules could leave it vulnerable to a sudden crippling turnaround in the deficit and a rapid accumulation of debt. The SGP points to what may be necessary, but are certainly not sufficient, conditions for sustainable public finance policies.

In Ireland’s case, the dependence on sources of tax revenue such as stamp duties, capital gains tax and corporation profits tax made total tax revenues dependent to an exceptional degree on profitable economic conditions, high transactions in the property market and asset price appreciation (Chart 2). All tax systems tend to do better in such conditions, but the elasticity of tax revenue response to a return to more normal conditions was exceptionally high in Ireland. Admittedly, tax revenues have slumped, and expenditure has soared in all of the countries affected by the global downturn of 2008-9 and the slow recovery that is now in process. (The taxes and spending programmes that behave in this way are normally seen as stabilizing, and indeed are called the automatic stabilizers.) But the deterioration in Ireland’s fiscal position has been worse than almost any other country. Partly this reflects the fact that Ireland’s home-grown property bubble had risen further and so has fallen further than most (Chart 3 – which presents a schematic picture of how GDP might have evolved absent our local bubble). But it also reflects the heightened elasticity of the tax receipts.

(Interestingly, when you plot tax and current expenditure as a percentage of GDP, you might miss this (Chart 4), masked as it is by the scale of the collapse in GDP – tax receipts fell even faster despite big increases in rates and base especially in income tax as the Government tried to make up for the shortfalls elsewhere.)

That is one of the main reasons why Ireland’s exposed fiscal position is so exceptional. The Government has been struggling with this evolving situation for the past two and a half years and it will be the subject of intensified corrective action to be set out in detail in the promised 4 year fiscal programme to be announced shortly.

How could one use accounting numbers to improve the usefulness of the SGP for ensuring sustainable public finances? I have some ideas on this to which I will return after I speak a little about accounting and the Irish banks.

A full account of the government finances is not complete in Ireland without considering the banks. Over the past year or more there has been intense interest in figuring out what will be the total cost of the losses incurred in the property bubble. Indeed this has been a focus of attention worldwide, with for example the IMF’s World Economic Outlook report devoting attention to the issue in each of its five semi-annual reports between April 2008 and April 2010. Interestingly, the estimated losses for the global financial system from this source have oscillated, growing from the first estimate of US$1.2 trillion to reach a maximum of $4.1 trillion before falling back to the latest estimate of $2.3 trillion (Chart 5). Clearly these estimates are not easy to pin down. This is not an exact science.

Although it is not an exact science, estimating expected losses is an important ingredient in determining what additional capital banks require to be able to withstand any future unexpected losses.

(From the prudential regulator’s point of view, it would be better if a realistic provision for expected losses were made and, in effect, subtracted from capital. Accounting rules and practice focus instead on losses that have already been incurred).

In times of crisis, incurred losses as reported in accounts tend to lag expected losses. Indeed, this is a matter on which the accounting profession really needs to reflect. Some of my colleagues feel that I should challenge the accounting profession for not having looked sufficiently far beyond the balance sheet date in providing their audit opinion. Arguably, in a rapidly deteriorating environment accountants should have been more concerned with post-balance sheet events concerns, and could have done better in seeing the need for expected loss forecasts and the amplitude of the potential property price movements in assessing the going concern assumption.

Whatever about that, if a bank is to retain the confidence of the market it needs to have enough accounting capital to absorb any additional expected losses and still remain well capitalized. This was the main goal of the exercise which the Central Bank embarked upon about a year ago and which has come to be known as the Prudential Capital Adequacy Review (PCAR). This required estimates of both NAMA-bound loans and the rest of the portfolios.

Estimates of expected non-NAMA loan losses for the two listed banks in NAMA were prepared by the Central Bank following extensive discussions with the banks. Much of the losses result from the NAMA haircuts; the associated loan purchases by NAMA crystallize those losses and trigger a formal recapitalization requirement for the participating banks. Accordingly, the additional required capital amounts to be sought by the banks from the market were announced to coincide with the first NAMA haircut announcement. At that stage NAMA valuations were available only for the largest 10 exposures. At 50%, the average haircut on this first tranche was much higher than had been expected earlier (a 30% indicative haircut had been officially mentioned in mid-September 2009). Nevertheless, these haircuts were read across to the rest of the NAMA portfolio in order to arrive at the minimum required capital. In the event – as was always possible – the second tranche haircuts, not fully announced until mid- August 2010, were even higher.

At this point, it became clear that a long-drawn out drip-feed process of NAMA tranche announcements based on each successive size-tranche was in prospect. In order to avoid this, NAMA agreed to conduct a sample review of the full range of tranches.

The resulting estimated all-tranche NAMA haircut for each bank was the basis of the end-September 2010 announcement which led to an increase in the required capital of Allied Irish Banks (AIB), and to confirming the judgment that AIB would not be able to raise all of the needed capital from the market.

The costs of Anglo Irish Bank (Anglo) depended even more heavily (than for the two big banks) on its NAMA haircuts, given the relatively high fraction of its portfolio being transferred. On the other hand, not being in the market to raise capital, an official judgment on the expected losses for its non-NAMA book was not needed until the final decision was taken by Government on that bank’s future. Nevertheless, it was important to acknowledge that the surprisingly high first tranche haircut for Anglo, if read across to the rest of the NAMA-bound book, already at that point implied an astonishing figure of at least €22 billion for the net long-term fiscal cost to the State, given the relatively aggressive provisioning already made then by the new management on the non-NAMA book. As is well-known, NAMA’s estimated all-tranche haircut for Anglo has somewhat increased this. In addition, the management’s evolving estimates of the prospective loan losses embedded in the non-NAMA book have also been increasing, leading to our recently published estimate of the additional capital required to capitalize the resolution and funding entities into which Anglo is to be resolved. (By the way, as accounting specialists you will recognize that required capital is different to net cost: if the expected loss projections prove to be accurate, the net long-term fiscal cost will be lower than the required capital sum to the extent that the tender offer on subordinated debt is successful and to the extent that residual capital remains at an eventual wind-up.)

To summarize, then, four main components have been key to arriving at a total cost figure for the losses created by the bubble, and exposed by its bursting: NAMA’s first tranche haircuts and the Central Bank’s non-NAMA expected loss estimates for the main banks (both published in March); NAMA’s estimated all-tranche haircuts and the Central Bank’s required capital estimate for Anglo, which takes account of the management’s revised non-NAMA expected losses (both as of end-September).

There are a few loose-ends of course:

  • For the smaller loans hitherto intended for NAMA but now staying with the banks, a provision equal to the average all-tranche NAMA haircut for that bank has been applied. This needs to be verified more thoroughly.
  • The final NAMA haircut, following the rigorous procedure mandated in law and required by the EU Commission, could deviate somewhat from the current estimate.

These could go either way, but in any case are relatively small elements. The current figures are as precise as can reasonably be expected at this stage.

For the future, the Central Bank will continue to conduct PCAR exercises on an annual basis. This will allow capital requirements to be kept consistent with evolving information on future economic and financial developments, including the performance of the residential mortgage portfolio. Progressively, these exercises will also look forward to taking account of the emerging Basel 3 capital rules.

Much has been learned through this process. Indeed, the reality is that traditional bank accounting practices were not really geared to deal with (i) assets that are so uncertain (and long-lived); (ii) an economy that has tanked so deep; (iii) future economic prospects that are so uncertain. But as the banks become increasingly adept at assembling the underlying information needed, so the PCAR figures will become progressively more reliable. Already we see differences with some bank managements taking a more aggressive approach to provisioning leaning more towards the expected loss idea than to the old-style interpretation of incurred loss. It may be no coincidence that new brooms have been to the fore here.

Let me return finally to the question of the Government accounts. I do not want to speak today about the necessary adjustments or to qualify any of the percentages that have been publicly discussed in regard to the future budgetary profile.

Anything that is unsustainable will stop. A sequence of double digit government deficits is certainly no exception to this rule. Absent a managed reprogramming of the fiscal adjustment, market forces will take over and put a stop to the unsustainable increase in the debt ratio.

What policy can – and I am hopeful will – achieve is that the adjustment will, not least by rolling back unaffordable excesses, be accomplished in a manner that is much less damaging to the economy and society. This means in particular ensuring that the adjustment helps as far as possible to restore lost cost competitiveness and enhance growth potential, while aiming at a fair distribution of the burden.

While the direct effects of the bank recapitalization have caused a sharp jump in the debt ratio, the dynamic unsustainability does not arise mainly because of this, as can be seen in Chart 6 which decomposes the changing debt-to GDP ratio of the Government into what is attributable to (i) the bank recapitalization (including interest), (ii) the fall in GDP and (iii) the rest of the deficit. The third element represents the largest – and growing – share.

(Servicing of the bank-related debt amounts to about 1 per cent of GDP – a sizable block, but only about a tenth of the prospective 2011 deficit.)

That’s why a sharp fiscal adjustment would have been needed anyway – even if the Anglo and Nationwide had not lost quite so much money.

There is a further link between bank accounting and government accounting which, I believe, deserves more attention as policymakers in Europe continue to develop new governance arrangements around the SGP. A leaf should be drawn out of the agreed international approach to bank capital and the distinction which I have alluded to between expected and unexpected losses. Thus, the fiscal position should not be considered merely in terms of the actual deficit (constrained to a 3 per cent target at present) but should also take account of the potential volatility of the deficit. A country which, like Ireland in the last decade, depends on volatile sources of receipt would be required to aim at a better fiscal performance – a lower deficit or even a surplus. A risk-adjusted deficit figure could be computed, just as bank regulators impose a risk-weighted capital adequacy requirement. Such a rule could insulate countries from the sudden emergence of a structural deficit. Even if such an idea might take time to command agreement at EU level, despite current discussions towards an independent fiscal review body, it should certainly be factored in to medium term fiscal planning at national level, to be brought into play when the budgetary accounts have been brought fully under control again in a few years time.