14 October 2011

Rerunning 2011 EU-Wide Bank Stress Test: Who Are We Fooling?

Dexia (that was one of the largest banks in Europe) comfortably passed the 2011 EU-wide Bank Stress Test; in fact, it came out from the stress test as one of the best capitalised banks. Barely three months later it is basically blown up and being sold in pieces (and, by the way, this is reality, not an imaginary stress scenario). Once again, the stress testing exercise turned out to be a joke (remember that last year these were the Irish banks that failed after successfully passing the 2010 EU-wide Bank Stress Test).

On Tuesday October 11th, the outgoing European Central Bank (ECB) president Jean-Claude Trichet warned that the European financial crisis had become "systemic" and demanded swift action.  (The Guardian, Wednesday 12 October 2011) This obviously because of the linkages between the Europe’s troubling governments and the far too leveraged banking sector (plus outside pressures to do something). So recapitalisation of the European banks is being seen as part of a wider scheme to resolve the sovereign debt crisis in Europe. How much additional capital is needed and which bank should receive (or accept) how much capital? The answer is that this should be determined based on a tougher capital test, i.e. by one more stress test this year!

Since there is no time for a full-fledged stress test, the second 2011 EU-wide Bank Stress Test is being performed based on the first one. The European Banking Authority (EBA) has asked banks to submit some additional data <See the template that was published by Reuters on October 11th: link.reuters.com/puc44s>. As reported by Reuters, the current test is expected to mark peripheral eurozone debt to market prices; in addition, the EBA wants banks to hold a minimum core Tier One capital ratio of at least 7 percent under a recession scenario (previously: 5 percent).

Now analysts are proposing numbers of how much additional capital the banks are going to need: is it 100 billion euros, 200 billion euros, 500 billion euros... “Breakingviews Euro zone bank Stress test” for example looks a pretty cool tool to play with. Credit Suisse AG analysts have calculated that at least 66 of Europe’s biggest banks would fail a revised EU stress test and need to raise about 220 billion euros; out of this, Royal Bank of Scotland Group Plc (RBS), Deutsche Bank AG and BNP Paribas (BNP) SA would need the most, a combined total of about 47 billion euros; Societe Generale (GLE) SA and Barclays Plc (BARC) would each need about 13 billion euros of fresh capital. (Source: Bloomberg, 13 October 2011)

But EBA’s officials, eager analysts in Credit Suisse, in Reuters and everywhere else, just hold on a bit: Do you really believe that a simple rerun of a misleading / failed stress test is a reliable basis for deciding which bank and how much capital is going to need? If yes, then I should say that I do not share your opinion. If not, then... who are we fooling when coming out with seemingly precise but in reality meaningless numbers? Ourselves? Investors? Tax payers?

I have discussed the issues related to this stress test earlier in my blog (and those closely involved into the stress testing exercise must know them even too well: subjectivity of internally assigned risk parameters, illusory part of the Core Tier 1 capital ratio etc.), see for example the articles “2011 EU-Wide Bank Stress Test: Scenario May Not Be Main Issue” and “Do We Need Bank Stress Testing At All?” A simple rerun of the test with modified assumptions for sovereign stress will ignore a significant part of them. This is for one thing.

Secondly (even more importantly), remember that banks like Dexia have gone broke and interbank markets may easily freeze because banks’ assets and liabilities are not properly valued (or are suspected to be improperly valued by the other financial market participants). One bank does not believe what the other bank is saying (and too often rightly so).  To illustrate this, consider the following.

In its 3 October 2011 press release, while announcing its request for help, Dexia’s Board of Directors is stating that: “However, in the current environment, the size of the non-strategic asset portfolio (so-called legacy) impacts the Group structurally despite the good credit quality of its assets.” Read: “Our legacy assets are of good quality, only no one does believe it. So we have run out of money.”

What is actually needed before any kind of stress tests is honest valuation of the banks’ assets and liabilities. Yes, basically the EBA (or whatever other authority) should ask for fair and comparable (i.e. based on the very same valuation rules and -models) presentation of the banks’ financial statements and make them publicly available. Only this can be the starting point of restoring confidence. We need to know, how large holes the banks have in their balance sheets today – not in case of some kind of hypothetical scenario.
Seems too trivial? Let me list some ways of creative accounting and other tricks that the banks are or can be using:
* Repo 105 and similar transactions to “reduce” leverage and “improve” liquidity
* Misleading presentation of Core Tier 1 and Tier 1 capital to demonstrate stronger capitalisation
* (Mis)use of internal models for reducing risk-weighted assets and demonstrating stronger capital ratios
* Artificially low loan loss provisions to “increase” profits or “reduce” losses
* Reclassification of financial assets from fair value accounting category to amortised cost category in order to “increase” profits or “reduce” losses
* Basing asset valuations to super aggressive forecasts to boost the values of the assets, particularly in what you would call model businesses
* Reducing the value of own debt (and hence narrowing losses, “reducing” leverage and “improving” liquidity) as the market value of that debt declines
* Boosting interest income via restructuring troubled loans and adding accrued (but not yet received) interest of the troubled loans to the principal amount of the restructured loan
* Zombie securitisation
* Doing repo so that one don’t have to sell and don’t have to take the loss on many of the assets upfront
Note that most of the above is perfectly legal.

Without a fair and comparable representation of the banks’ assets and liabilities, injecting 200 billion euros or 300 billion euros into the banking sector is just another temporary relief (thanks to the very fact that the banking sector will have more capital). First, the distribution of bailout money between the banks will not be optimal. Secondly, the very problem of lacking credibility is still not being addressed.

Sure, the task is challenging to say the least. Indeed, what is the money value of an asset or a liability in a world where the value of money itself is an illusion (read also the posts in the categories “Basics” and “Money”, starting from the “Money Creation – Basic Principle”)?

1 comment:

  1. As you wrote "Dexia (that was one of the largest banks in Europe) comfortably passed the 2011 EU-wide Bank Stress Test; in fact, it came out from the stress test as one of the best capitalised banks. Barely three months later it is basically blown up and being sold in pieces"

    Says it all really doesn't it.

    It would be fair to say that stress tests are missing something ~ correlation and dependency analysis. Most banks stress tests are too linear too far away from the multivariate copula.

    Stress tests need to capture multiple impacts on funding liquidity, downgrades, runs on deposits, drops in share price and the inability to balance ALM gaps with short term roll overs.

    Banks can't simply add up these exposures separately in the backdrop of tight liquidity and market volatility because each element compounds other factors on the balance sheet.