The test results are in and markets aren’t happy. Far from assuring investors, the July stress test results for European banks prompted a downward spiral of bank stock prices (Figure 1). All the indicators are down for the year; for most nations, last month’s route accounts for the bulk of the annual drop. We’ve not seen anything like this since the Lehman collapse.
Figure 1. Bank stock performance as of 11 August 2011
Source: UBS Global Bank Valuation. The graph compares the performance of banks stocks grouped by country, on the last year and on the last month, i.e. since the publishing of the stress-test results.1
Valuations drop despite favourable stress-test scores
These market moves contrast sharply with the immediate results of the 2011 tests--in particular with the satisfactory result that was found in terms of banks' capital adequacy.
- Only 8 banks failed the 5% hurdle for Core Tier 1 capital over the two-year horizon;
- 16 banks fell within the quite safe range between 5% and 6%.
These favourable outcomes are due in part to the preemptive capital increases (more than €50 billion) that have arrived in recent months--generally under the pressure of national bank regulators.
It is worth remembering that the macro hypotheses underlying the 2011 stress tests are more severe than the previous year’s and they didn’t take account of the stronger-than-expected economic recovery of recent months. This would appear to make the markets' reaction even more difficult to explain.
Were the stress tests too easy?
Some accused the tests of not being severe enough. A common theme is the observation that they skipped over banks' sovereign debt risk to banks' trading and banking books. This point, however, misses the main innovation of the 2011 stress tests.
This year’s tests released an unprecedented amount of data on individual bank exposure--including ‘exposure at default’ data (one of the factors that determine banks’ regulatory capital requirements). The new data means that the market has all the information they need to roll their own stress tests, i.e. to evaluate each bank’s exposure to private and sovereign risks.
This is a major step forward. It is clear evidence that the newly-born European Banking Authority managed to overcome the resistance coming from banks and national regulators. It is proving to be a fully-fledged European authority, not just a committee of national regulators prone to compromise (as was the case for its predecessor, the Committee of European Banking Supervisors, or CEBS).
Private “stress tests” based on the new data
Using the newly available data, a number of analysts investigated banks’ health under tough assumptions on possible sovereign debt losses. These are worth considering. Goldman Sachs (2011) estimates that:
- Applying a haircut to the debt of Greece, Ireland, and Portugal would produce a bank capital shortfall of €25.9 billion relative to the 5% Core Tier 1 capital benchmark.
- Applying a further 10% haircut on the Italian and Spanish securities, the shortfall would rise to €29.8 billion.
Barclays (2011) estimates that only four banks of the top 30 would fail the 5% hurdle when applying mark-to-market also to the banking book.
Of course, it must be taken into account that losses (and capital shortfalls) would not be evenly distributed across the European banks. They would fall mainly on the banks of the defaulting countries. Nevertheless the key point is that the shortfalls--and thus the necessary injection of capital in the event of haircuts--are smaller than the ones that have already taken place over the last 18 months.
Given these results, the negative reaction of the markets (reinforced later on by bad news on the American debt and the fragility of the world’s economic recovery) seems difficult to explain.
It’s the bottom line, my dear
Stress tests also provided data on banks’ "net interest margin", i.e total interest revenues minus total interest costs, funding costs and net profits after taxes. If we look at this part of the exercise, the situation changes significantly (see Figure 2).
- In 12 of the 30 most important European banks, the 2012 “adverse scenario” produces actual losses; for another 5, the result is only very small profits.
- By contrast, in 2010 only a few banks recorded a loss (two Irish and two UK).
- Top names such as RBS, the Italian banks, and the German banks (with the notable exception of Deutsche Bank) would record significant losses.
- French banks (excluding SocGen) would see their profits drop almost to zero.
- Only Spanish banks and Barclays would still record significant profits.
Figure 2. Net profit after taxes (NPAT) and Core Tier 1 capital ratios in 2012 under the “adverse scenario”.
Source: European Banking Authority
Of course, losses and provisions of the adverse scenario explain a large part of this difference. But a great lot is explained by the profitability of the traditional intermediation, measured by net interest margins.
- In the adverse scenario, only Deutsche Bank foresees an increase of its margin (by a surprising +22%).
- Other banks would have to face a decrease of their margins, which could reach worrying levels for the Irish banks (-75% and -52% for Allied Irish and Bank of Ireland, respectively).
In an industry where fixed costs are high and quite rigid, a reduction in the absolute value of the profitability of the traditional intermediation could threaten, even permanently, the underlying profitability of the banks and, therefore, their ability to attract capital in the long run.
It’s the funding costs
The main reason behind the reduction of net interest margins lies in the increase in funding costs. It is important to stress that in the present phase of the financial crisis, sovereign risks and banking risks are strictly linked together. As IMF (2011) and BIS (2011) have shown, increased funding costs is the main transmission mechanism between the perception of higher sovereign risks (and a corresponding increase of the spreads vis-à-vis the typical risk-free asset in Europe, the German Bund) and the increase in interest rates on banks funding, particularly through bonds which are nowadays a significant portion of their liabilities.2
This paints a worrying scenario of a generalised shock on banks’ interest margins and therefore on their core profitability, linked to the funding costs.
- The stress tests have revealed that at the end of 2010, the 90 banks had €8.2 trillion of wholesale/interbank funding (more or less equal to the Eurozone’s GDP).
- €4.8 trillion of this matures in the next 24 months.
- €2.7 trillion of it matures in 3 months.3
As the last Financial Stability Report from the ECB warns: “Bank funding risk has been one of the most important sources of banking sector vulnerability throughout the financial crisis. Indeed, one of the notable implications of the ongoing sovereign debt crisis was the intensification of this risk” (ECB 2011 p. 92).
Strong capital and shaky profits is thus the message from the stress test
In other words, over and above the prospect of losses on banks’ portfolios, there is also the danger of a significant shock on their core profitability.
- We can therefore conclude that if the markets are bearish in the valuation of bank stocks, they have a point.
Undoubtedly, valuations well below the book value seem too pessimistic and even foolish, but as Polonius would comment on Prince Hamlet, “Though this be madness, yet there is method in it”.
If we remember the results of the research from Goldman Sachs and Barclays, we can also say that European banks could reasonably absorb a one-off shock from a default of the three peripheral countries, but are worst off in a scenario of muddling-through where sovereign risks remain high and therefore market spreads bring an unbearable cost not only for the government but also for banks’ liabilities.
The failed strategy to fix the system
The strategy followed by governments and regulators since the inception of the financial crisis has hinged on two pillars:
- Creating a favourable environment for banks’ profitability (based on abundant liquidity and very low interest rates), while
- increasing the capital levels (to allow future losses to be absorbed).
The growing problem of the sovereign deficits was tackled separately by imposing fiscal discipline in proportion to national debt burdens. This strategy has backfired.
The failure is caused by simple facts: the two problems are two sides of the same coin, and the solution was based on overly optimistic assumptions on public debts sustainability. The problems were viewed as a liquidity problem not a solvency problem.
Explaining periphery-to-core contagion
Given the facts noted above on the banks’ health, the spread of contagion is easily understood. The peripheral countries' debt could be restructured without significant losses for the banking system. However, concern has spread to core European countries such as Italy and Spain creating the worst of all possible worlds for financial institutions. Uncertainty on their assets’ value is compounded by the uncertainty on their core profitability.
The stubborn refusal to admit a solvency problem for small European countries (and small European banks) has kept interest rates high. These are levels that run the risk of making sovereign default a self-fulfilling prophecy, thus creating the danger of a new, unsustainable shock to the financial system.
The policy response has failed the market test
All policy measures worked out in Brussels have failed the market test. One can argue that the drop in bank stock prices was not triggered by the result of the stress tests. Rather, it was caused by the dismal compromise worked out in Brussels the following week (21 July 2011).
- Now, the haircuts implied in Barclays’ and Goldman Sachs’s exercise are probably insufficient and losses could be even higher.
- We now know beyond any reasonable doubt that the plan to stabilise the banking system through recapitalisation and low market yields is not apt to the difficult times we are living.
It is time to call a solvency problem by its name.
- We should restructure the debts of Greece, Portugal and Ireland to levels that make the burden reasonably sustainable in the long run.
- We should protect the countries such as Italy and Spain which suffer only from the present poor growth.
- Bail-in creditors, including banks, and ring-fencing would come together.
This restructuring is still sustainable by the European banking system and it is a cost that is worth paying, even if it could mean the (hopefully temporary) nationalisation of the banking system of the peripheral countries.
Every financial crisis has been stopped only when the measures taken have been severe enough to reverse the market expectations and increase the downside risk for speculators. We need to return to a world where the market attaches no risk, or a low risk, to securities issued by major governments and the major banks.
In this framework, even a small amount of bank capital is adequate and more ambitious targets in terms of quantity and quality of capital can be gradually achieved along with the improvement of the economic outlook.
This time is not different and the more the situation worsens, the more the shaky situation of the European banks threatens the entire European building.
Barclays Capital (2011), “Life Goes On. The European Stress Test”, Equity Research, 18 July 2011.
BIS (2011), “The Impact of Sovereign Credit Risk on Bank Funding Conditions”, Committee on the Global Financial System, Basel, July 2011.
EBA (2011). EU-Wide Stress Test. Aggregate Report, European Banking Authority, 15 July 2011.
ECB, Financial Stability Review, June 2011.
Goldman Sachs (2011), “Stress Test II: Banks not buckling under EBA’s Stress; Focus moving to connectivity”, Goldman Sachs Equity Research, 18 July.
IMF (2011), Global Financial Stability Report, April 2011.
1 UBS Research, Global Bank Valuations, 11 August 2011.
2 The stress test data reveal that in 2010 the (unweighted) average for the funding costs of the 30 banks was 179,5 basis points, with a wide dispersion, ranging from a minimum below 100 for Barclays and Deutsche Bank to a maximum of 280 for Bayerische Landesbanken. Under the hypotheses of the adverse scenario, the funding costs are expected to jump to 331,1 basis points (an increase of 84.4%), ranging from a minimum of 168 for Nordea Bank to a maximum of 446 for Bayerische Landesbanken. It is interesting to note that the dispersion of funding costs narrows: the coefficient of variation (the standard deviation divided by the mean) decreases from 0.26 to 0.17.
3 EBA (2011).