He observed that the markets were reluctant to lend to Ireland because of concern with the tail risk of the assets in the Irish banking system, including NAMA.
As regular readers of this blog know, it is only by disclosing current asset-level data to all market participants that the market can judge the tail-risk of these assets.
Unfortunately, as Mr. Honohan discusses, the Irish government did not choose to disclose current asset-level data. Instead, it chose to run stress tests.
As discussed in a previous post, the fundamental problem with stress tests is that the Irish government is an obstacle to markets functioning properly. In this case, they are blocking the disclosure of the current asset-level data that the market wants from reaching the market. The Irish government is doing this by running stress tests.
This will be at least the fourth time that the Irish government has tried to find a level of disclosure that restores market confidence without disclosing current asset-level data. The first couple of times involved buying the "bad" assets. The last couple of times have involved stress tests.
In an effort to make these stress tests more credible, the Irish government has hired BlackRock, Barclays and Boston Consulting. However, these firms are constrained by the terms of the Irish bailout to find a maximum of 35 billion euros need to cure the problem. This constraint immediately discredits the analysis.
What is more important is the fact that the only way to convince investors is to provide them with the current asset-level data and let them analyze it for themselves. What their analysis shows is what they will believe in.
...As seen from the rest of the world, and in particular from the financial markets, the key risks now facing the Irish economy lie in its financial balance sheet. To oversimplify, Ireland faces two main problems in this respect. First, it has too much debt, public and private. Second, there is a market perception of significant tail risk to the debt, not least that part that relates to the banks. It is these two problems, which are obviously interlinked, and which have both been growing in the course of the past year, that have led to the market’s reluctance to provide continuing funding at reasonable rates of interest.
... If I am right in saying that the market’s reluctance to finance Ireland is down to the exposure to tail risks that they fear, then restoring Ireland’s access to market funding needs to address those risks. There are three possible routes. First is to improve the quality and transparency of relevant information so that those parts of the perceived risk that are based on inadequate information will be dispelled. Second, to reduce vulnerabilities through mitigants – notably reducing the leverage of banks, reducing the fiscal deficit and thus slowing the accumulation of government debt. All means of restarting growth of output and employment on a sustainable basis fall into this category. That would include sustaining the recovery, under way since 2007, in wage competitiveness in international markets, as well as strengthening other elements of international competitiveness. Most of the key components of the years of “Celtic Tiger” success are still in place: clearly none should be neglected or dismantled at this juncture.
... If so, we must work with what we have, and that is the policy strategy worked out for the IMF-EU Programme. It has two main elements, which essentially continue and deepen existing policies.
First, there is phased deficit reduction projected to bring Government debt under control and declining by 2014. After the period of political uncertainty engendered by the run-up to the General Election, the new Government has reaffirmed the fiscal objectives of the Programme and specifically to the 2011 and 2012 budgetary numbers. This is key to reducing risk and restoring confidence, but I will say no more about it here.
Second, accelerated resolution of the banking crisis for which we would have hoped to turn the corner already, but has proved to be on a scale which has challenged available policy tools. The banking strategy thus involves further recapitalisation and accelerated downsizing.
The recapitalisation of the banks will take place following the completion of the 2011 stress tests which, in line with the Programme commitments, will be much more granular, more explicitly transparent, and tougher, than those of 2010.
Last year’s task was not just a question of assessing future losses under a speculative stress scenario; instead we were still trying to evaluate the losses already embedded in the existing portfolio resulting from the property price collapse that had already occurred, especially for development property.
In normal times, a well-run bank with good loan appraisal procedures will be able to forecast its likely loan losses fairly accurately, especially if the likely macroeconomic prospects can be pinned-down. But when it becomes clear that banks have made substantial underwriting errors across a wide range of their portfolio to the extent that began to become evident in late 2008 and early 2009, the range of uncertainty balloons out. Standard valuation practices that work for properly underwritten loans in normal times are ineffective in such an environment in predicting the enormous systemic losses that can occur. Thus, over the past couple of years, a point estimate – a single number – has been more misleading than helpful in terms of understanding the range of loss possibilities.
Usually the validity of stress tests is revealed only over a period of time, but in our case, within months of publication, one of the underlying assumptions of the 2010 stress tests was unluckily undermined. We had to go public on them part-way through the big programme of loan purchases by the State property company NAMA. These loan purchases crystallised those embedded bank losses immediately, but the prices at which the purchases would be made were calculated in a very elaborate and time-consuming way mandated by Competition Law. The stress-test’s seemingly aggressive assumption that the seemingly severe haircuts in the first tranche of loan purchases would read through to the lower tranches also proved to be wrong. With hindsight it would have been great to have built in more insulation by providing for even tougher haircuts; but at the time it would have seemed an arbitrary and hard to justify exercise of regulatory powers.
This year, we are determined to present a stress test that will not only be more convincing to the market, by providing much more granular detail, but also be better insulated against surprises by using aggressive stress assumptions and modelling.
The emphasis is clearly on determining appropriate capital levels to meet stresses. The higher capital required will be driven by this rather than from any increase in expected losses: to be sure the somewhat weaker economic growth projections now available would imply some increase in expected losses, but it is the stress scenario that is the focus of the present exercise.
We will seek to largely free the tests of any excessive expectations from Irish exceptionalism in loan-loss recoveries (over the years very few Irish residential mortgages have been foreclosed in downturns by comparison with the experience in the UK and US, for example: with the help of external consultants we are referencing experience in those countries – though not slavishly so – in the analysis that will lead us to choose new tougher capital levels for the banks sufficient to convince the markets. And less reliance will be made on extrapolation from recent experience.