Review of the Irish stress tests

Just like Judge Moriarty’s 2400-page report last week which may cost the State €250m and deserves to be read cover to cover, so also does the Financial Measures Programme published yesterday by the Central Bank of Ireland (CBI). It represents the output from an exercise reported to have cost €20m and having now studied it, I can offer the opinion that it is a text worthy of study by all; even if this is not your area, there’s a decent five-page glossary included that might help anyone who wants to understand the whole banking crisis. And at less than 100 pages, it’s a pamphlet compared to the Moriarty tomes. There is a fair smattering of presentational errors, mind eg table headings on page 31.

The key question probably is “is the CBI report any good?” It uses work from BlackRock Solutions, project managed and “challenged” by the Boston Consulting Group. It was peer-reviewed by staff from the Italian and French central banks as well as named and un-named experts and consultants in Ireland – the named third parties are Deloitte, Mazars, Ernst and Young, Clayton Euro Risk Management, Situs, Arthur Cox, Matheson Ormsby Prentice. Funnily enough Barclays Capital doesn’t get any mention. But based on the professionals involved,  it probably should be good but for myself, I found the report unconvincing but that was mainly through the non-disclosure of information (what interest rate assumptions are used, for example). A key concern on here is property values, and I am not sure that a 59% peak-to-trough fall for residential and a 69% peak-to-trough fall for commercial, both in Ireland, is sufficient to cover “even extreme and improbable losses”. That said, who amongst us has a crystal ball. But still.

(1) According to the CBI, the stress (=adverse) loan losses are “not considered likely to materialise; they are merely an input designed to ensure the associated capital requirements are fully convincing to the market as being sufficient to cover extreme and improbable losses”

(2) BlackRock Solutions is responsible for the loan-loss estimates. Other inputs from elsewhere are also used in the calculations.  Many of the inputs like interest rates, probability of default, losses given default just aren’t shown. With respect to interest rates, there is a statement that the rates used in the European Banking Authority stress tests which are now ongoing were used (1.5% in 2011 and 1.8% in 2012) – however those rates just go to 2012 and are base rates as opposed to retail rates. Compare this with the UK stress test documentation two weeks ago, and for me, I’m left scratching my head as to whether our own tests are sufficiently transparent.

(3) The aim of the programme, of which these stress tests form part, is to place the Irish banking system in a position where it can function without a reliance on Irish and other state supports (including the IMF presumably). Yesterday’s publication is presumably the start of the process because for me, what is contained in the publication yesterday is not sufficient to convince risk investors.

(4) When reading the BlackRock loss estimates you should bear in mind that they exclude existing provisions for losses.

(5) BlackRock’s modelling extends out to 30 years but the results are really only examined over the next 33 months to the end of 2013

(6)  The CBI engages in some jiggery pokery with its introduction of a new banking term “additional capital buffer”. It might just as easily stated the true Core Tier 1 capital incorporating the “additional capital buffer”.

(7) The four financial institutions (hereafter called “banks” even though EBS is a building society and Irish Life and Permanent is a bancassurer) are forecast to make operating profits of €3.9bn in the three years – 2011, 2012 and 2013. Operating profits exclude loan losses and any losses that arise on deleveraging.

(8) The four banks are forecast to lose €13.2bn on selling assets in the next three years. This is a phenomenal cost – it is practically €100m per week, every week for the next three years. And this loss is unnecessary in the sense that it will only arise because the banks are being forced by the ECB/EU/IMF to sell assets now.

(9) The deleveraging plans are a bit of a paradox. On one hand there is a commitment to sell assets in the next three years. On the other hand there is a commitment to avoid fire-sales. What happens if macroeconomic conditions means that the irresistible force meets the immoveable object?

(10) The CBI comes up with a form of words to deny the baseline scenario is a forecast – “In fact, given uncertainties in the current climate, it is improbable that either the base or stress scenario will prove to be accurate across the macroeconomic indicators”. Hmmm, I’m still not convinced.

(11) The Loan to Deposit Ratio (LDR) was 180% at the end of 2010. The target at the end of 21013 is 122.5% “agreed with the External Partners” This is the first time that I have seen this published.

(12) At the end of 2010 the four banks had gross (before provisions for losses) loan balances of €274bn

(13) Page 22 says that BlackRock used forward assumptions on house prices and interest rates. But no-where in this document is there any detailed information on the interest rate assumptions.

(14) BlackRock assumed that house repossession levels would converge with the UK, not Nevada as reported in the media yesterday. Again, it is not clear what those UK levels are – the distressed ones in the early 1990s or ones prevalent today.

(15) Derivatives and other non-loan balance sheet exposures. You might recall Pat Rabitte saying on RTE radio last year that he was concerned by rumours of multi-billion euro exposures at some banks. This review  seems to conclude that there is no exposure but this is what the sole paragraph in the entire 88-page report has to say on the subject (emphasis mine) : “In addition to forecasting loan losses, Black examined banks’ securities and derivatives portfolios.  BlackRock analysed a portfolio of €59.8bn in securities in 1,597 individual positions. The majority of the securities at the institutions were used as liquidity instruments for cash management purposes. AIB and BOI also supplied data on small portfolios of more illiquid assets held on their corporate balance sheet. Pricing discrepancies for all positions spot-checked were not material. BlackRock also performed a high-level review of individual derivatives positions. These instruments were almost exclusively of the type generally used to support hedging and funding programmes, consistent with the institutions’ stated objectives. In the time available, the objective was not to conduct a full review of positions. Instead the intent was to provide comfort on the nature of the portfolios by assessing their size and shape and conducting non-statistical spot checks. Spot checks revealed no material differences” For myself, I am not re-assured by “non-statistical” checks and it is also unclear who oversaw, or “challenged” this work.

(16) BlackRock is estimating lifetime losses on the €274m loan book of €28bn (baseline) and €40bn (adverse). Having read this report a couple of times I am still unclear how the CBI has arrived at a 3-year stress loss of €27bn but the CBI does say that it discounted BlackRock’s numbers by 31%. Or more specifically how did the CBI think that a €3bn of additional “contingent capital” for losses after 2013 was going to be adequate to cover €13bn of losses that BlackRock is projecting?

(17) Anglo and INBS get a footnote mention in appendix I. The conclusion is that existing loss estimates are adequate as supposedly verified by a PwC review in Q3/4, 2010 and an “Independent Consultancy” review in Q4, 2010 and applying the adverse stress tests to the INBS and Anglo loans books.  I would be sceptical. For example last year it was reported that one third (yes 33%) of INBS mortgages were in arrears compared with less than 5% for all banks. There has been consistent criticism of Anglo’s and INBS’s lending standards. So is it appropriate to apply the stress test results from the other four banks? I don’t know how exactly they have been applied but I remain sceptical.

So how good is it? Well, it will be the market that decides whether or not to start lending to Irish banks again. The announcements yesterday won’t in themselves make that happen, but will the implementation of the measures outlined – the deleveraging, the restructuring, the additional capital injections – allow our banks to wean themselves off central bank support? Time will tell. Was there any avoiding the stress testing? Actually there probably was, because the EBA is undertaking the similar work though there is an argument that the involvement of BlackRock and the detailed bottom-up testing was needed here. Could the stress tests have been improved? Probably, certainly the review of derivatives looks cursory, certainly there is a dearth of information and justification for the assumptions used. Like many things in life I suppose, time will tell.

UPDATE: 9th May, 2011. The Irish Independent reports that the cost of the stress test was €30m and that in addition to the three previously-identified service providers (BlackRock Solutions, the Boston Consulting Group and Barclays Capital),  seven other local companies were also engaged. The companies named are Ernst and Young, and law firms Arthur Cox and Matheson Ormsby Prentice. It has not been disclosed how much each firm was paid. What is concerning about the revelation of the involvement of these firms is the commitment that was apparently given in the IMF/EU Memorandum of Understanding that firms that had been involved previously with Irish banks would not be used in the stress tests as it was felt a fresh set of eyes might produce more credible and accurate results. The breakdown of the €30m bill will be of interest as it will indicate the degree of involvementof each firm. Did BlackRock bill €5m and Ernst and Young bill €10m?