For the second time in two years, the European Banking Authority (“EBA”) completed tests on European banks to demonstrate their “solvency” under conditions of “stress“.
The results have been over shadowed by other momentous events – the announcement by the European Union (“EU”) of a range of measures to deal with the European debt crisis. The tests remain highly relevant as the EU measures are unlikely to “resolve” the debt problems and European banks remain heavily exposed to losses. The risk of a European banking crisis remains.
Last year’s debut test was openly derided as the equivalent of conducting a crash test on a motor vehicle assuming that the car could never crash. Within months of the stress test, several Irish banks that had passed needed rescue by the EU and International Monetary Fund (“IMF”). Another major casualty was the credibility of European Banking regulators.
For the record, only 8 out of 90 European banks tested “failed” the 2011 test, meaning that they did not have the required 5% level of capital in a stressed scenario. Their capital needs were estimated at around Euro 2.5 billion. Another 16 banks were deemed vulnerable, being within 1% of the minimum required capital level. The numbers are eerily similar to that in the first stress test, which passed all but seven of the 91 banks tested identifying an aggregate capital shortfall of only Euro 3.5 billion.
In another similarity to the original stress tests, all those who “failed” were small banks. The 2011 tests did show that some large banks, including Deutsche Bank in Germany, UniCredit in Italy and Société Générale in France, had capital reserves close to the minimum level where they would have formally been required to raise more capital or reduce risk.
Highly stage managed and contrived, the latest stress test has done little to build confidence, instead exacerbating uncertainty and doubt.
The coverage of the stress tests is not comprehensive covering around 60-70% of the European banking sector, with many small banks being excluded. After the 2011 test, two small Danish banks, who had not been tested, failed. While their assets were minor, the ability of a small bank failure to affect other larger banks in a chain reaction remains a concern.
During the 2011, one bank – Helaba – withdrew, taking issue with the definition of capital. The level of coverage reduces the utility of the test and the value of the information disclosed.
The test focused on the effects of weaker economic growth, falls in stock and property prices, increases in funding costs and, most topically, losses on sovereign debt holdings. A well recognised problem is that the tests were really not “stressful” enough. The most significant deficiency was in relation to bank holdings of sovereign debt.
Where sovereign bonds were held in the banks’ “banking books” (generally long term, non traded investments), a credit rating downgrade of the issuer was specified – 4 notches for a country rated BBB, or lower, 2 notches for a country rated AA/A. For Greek bonds, this translated into a stress scenario of 15% loss. This was below the actual market losses of 50-60% implied by current market prices for Greek bonds or the 21% loss where banks roll over Greek debt under the new bailout package proposed. Similar problems exist with the Irish and Portuguese bonds where market prices imply losses around 30-40%.
Unlike the previous stress test, sovereign bonds held for sale in the bank’s “trading books” were included. The test implied losses from widening credit margins of around 20-30% for Greek bonds, with lower losses for Ireland and Portugal. For Spanish and Italian bonds, the losses were more modest, below the levels seen in recent trading.
The test results may significantly underestimate losses if Greece, Ireland and Portugal default and the Spain and Italy experience serious financial pressure. Traders joked that they would have “mark-to-market” profits if the stress test standards were used to revalue their holdings.
Other stress elements, while less controversial, were also insufficiently testing. The modest drop in growth and rise in unemployment contrasts with the sharp contraction in economic activity and employment seen in the peripheral European economies. Given the stated desire of the EU to bring public finances and budget deficits in line with guidelines, the risk of a long period of stagnation – low growth and high unemployment – and its effect on loan losses cannot be discounted.
The use of consistent parameters across Europe is questionable. Country-by-country credit data disclosed by the stress test reveals the fragility of some economies. For example, in Spain, British banks disclosed default rates of 8-9% on their credit portfolios. In Ireland, one British bank suffered defaults totalling Euro 13 billion on a total credit portfolio of Euro 64 billion – a rate of loss of around 20%. Standard stress conditions unadjusted for specific national conditions are analogous to achieving an acceptable average ambient temperature by placing one’s feet in the oven and head in the refrigerator.
The assumption of a 15% fall in stock prices may understate potential volatility. The 10-17% fall in property prices looks benign. For example, Spanish property prices remain elevated, well above trend. The overvaluation is exacerbated by unwillingness on the part of lenders to take possession of and sell assets securing loans, which would realise losses.
The interaction of the individual stress factors, the so-called “knock on effects“, are crucial. In the first phase of the global financial crisis, major problems were caused by negative feedback loops starting with loan losses, distressed banks, freezing up of money markets, cessation of lending and macro-economic effects (reduced growth, higher employment, slowdown in trade) which then set off new cycles of instability. The non-incorporation of “second order effects” is a serious deficiency.
Wrong Kind of Testing
A more egregious problem is that it was perhaps the wrong kind of testing on the wrong parties and issues.
Financial systems rely on the implicit support of national banking systems by the sovereign. As highlighted by the global financial crisis (“GFC”), financial institutions are “too big too fail.” This assumes the ability of sovereigns to support their banks, should it become necessary. The financial distress of many Euro-zone countries is well documented. These sovereigns are in no position to support their banks. A stress test of European banks is meaningless without a parallel assessment of the ability of the sovereign to support the bank.
This affects the access of banks to capital or liquidity. The stress tests merely identify the need for capital. Banks must raise capital from the capital market or governments, as was necessary during the first phase of the GFC. Given that a troubled bank may face difficulty accessing capital from investors, the ability of the governments to supply the necessary capital is crucial.
No one seriously believes that European banks need a mere Euro 3.5 billion of additional capital. Rating agency Standard & Poor’s estimates that European banks need around Euro 250 billion of additional capital to withstand a sharp economic downturn. Using a stricter 7% capital target and adjusting for losses on sovereign debt in banking book, JP Morgan research analysts believe that around 20 banks would need to raise capital. UK banks would have tor raise around Euro 25 billion, French banks Euro 20 billion and German banks Euros 14 billion. The ability of national governments to provide banks with the required additional capital remains unknown.
The stress tests also focused on the banks’ assets. The trigger for bank problems is usually the inability to finance themselves in the money markets – the “run on the bank“.
The funding difficulties of certain European banks are well known. Greek, Irish, Portuguese, Spanish, smaller Italian and smaller German banks currently have limited access to international money markets.
Covered bond issues, a traditionally reliable means of raising funds, have had to be cancelled. Some European banks have abandoned financing plans in the US and Asia in the face of nervousness in money markets.
Many banks, especially in the troubled peripheral sovereigns, are now dependent on funding from their own central bank or the European Central Bank (“ECB”). According to market estimates, the ECB has provided up to Euro 400-450 billion in financing, including Euro 240-250 billion to the bailout nations. Much of this financing is against collateral of uncertain value, being government bonds including those issued by the beleaguered sovereign. Some stronger national central banks, according to some studies, may have also advanced Euro 450 billion to other central banks within the Euro-zone to assist with financing banks with liquidity problems. The Bundesbank’s share of this financing is over Euro 300 billion.
The stress tests did not assess the banks’ ability to withstand continued funding pressures. The EBA glibly noted that it was “aware of the funding liquidity challenges in the current environment and national authorities are taking steps to extend maturities, increase buffers and develop contingency plans.”
Facing criticism about the rigour of the tests, the EBA pointed to the disclosure of financial information which enables banking analysts to conduct their own review. The volume of data is impressive – 91 banks; 3,200 data points per bank; 10 pages of disclosures per bank. Unfortunately, the information provided is filled with anomalies, mainly relating to aggregation of data and omissions.
For example, one British bank only disclosed the European credit exposures in its UK operation, omitting exposures held through other entities below the 5% EBA disclosure threshold. One Italian bank did not separately disclose Eastern European exposures, as they were held via their subsidiary in Austria.
The stress tests are revealing about EU and European banking sensibilities. The design of the sovereign bond tests highlights the underlying politics and semantics. The use of the rating agency methodology (4 notch downgrade) allowed the EBA avoided the use of the “banned” terms – “default” or “haircut” on sovereign bonds.
Europe seems unable to grasp the seriousness of the situation and deal honestly with its financial problems. There are media suggestions that the EBA preferred a stronger stress test but was constrained by political and banking pressures. There were even suggestions that national central banks failed to provide EBA with accurate data. The UK Telegraph quoted the EBA Chairman saying that: “figures given by the banks in some cases ‘materially’ changed after being challenged”.
Bankers resisted disclosure arguing that given the uncertainty in money markets, release of detailed information would exacerbate the crisis. The answer, argued bankers, was to reduce the level of detail significantly.
The stress test’s objectives were laudable: (1) to strengthen the financial system by forcing banks to hold adequate equity, and (2) to convince markets about the solvency of European banks. Unfortunately, the deep flaws of the stress test mean that they failed both objectives. The market’s judgement was quickly evident. Straight after the release of the test results, Italy’s Intesa Sanpaolo and Spain’s BBVA, who emerged from the stress test with core tier one capital of more than 8% – a top rated “pass“, saw their shares fall sharply.
When the expected European sovereign debt restructurings occur and the effects on the real economy become evident, European banks and regulators will still need to pass the only test that really matters. Whatever the stress tests say, the real test still lies ahead.
Satyajit Das is author of Extreme Money: The Masters of the Universe and the Cult of Risk (due to be published by Penguin Australia in August 2011)