Abstract. This paper is aimed at finding banks' destabilizing behaviors that explain why the impact of the crisis is so serious in the banking system. By comparing three German banks stuck in the crisis, I find that: I) the leverage is a common destabilizing factor and, ii) the banks were highly interconnected to other financial institutions and had a large maturity mismatch were more seriously affected by the crisis.

Peixin ZHANG
University of ParisWest Nanterre La Défense, EconomiX

LINK

Executive summary

The Committee of European Banking Supervisors (CEBS) was mandated by the ECOFIN of the European Council to conduct in cooperation with the European Central Bank (ECB), the European Commission and the EU national supervisory authorities a second EU-wide stress test exercise.

The overall objective of the 2010 exercise is to provide policy information for assessing the resilience of the EU banking system to possible adverse economic developments and to assess the ability of banks in the exercise to absorb possible shocks on credit and market risks, including sovereign risks.

The stress test has been conducted on a bank-by-bank basis and using bank’s specific data and supervisory information.

CEBS has coordinated the exercise and conducted extensive cross-checks over the results, which were submitted to a peer review and challenging process in order to ensure the consistency and comparability of the results. This report provides details on the scenarios, methodologies and aggregate results of the stress test exercise. Results of the individual banks and comments on follow-up actions, where needed, are provided by the banks participating in the exercise and/or their national supervisory authorities. The results are re-published by CEBS on its website.

National supervisory authorities routinely conduct stress testing exercises in their respective jurisdictions, both at system-wide and individual institutions’ levels, in order to assess potential risks facing the institutions individually and/or collectively. The CEBS exercise is intended to complement these national analyses by providing a coordinated assessment of European banks, using common scenarios and methodologies.

However, as with any stress test exercise, the results are not forecasts of expected outcomes, since the scenarios are designed as "what-if" situations reflecting extreme assumptions, which are therefore not very likely to materialise. Against this background, the aggregate results discussed in this report as well as the individual results presented by banks and/or national supervisory authorities, aim at supporting the supervisory assessment of the adequacy of capital of European banks, and should be interpreted with caution.

Sample of banks

The 2010 stress test exercise has been conducted on a sample of 91 European banks (The sample of the 2009 exercise was composed of 22 large cross-border banks.). In total national supervisory authorities from 20 EU Member States participated in the exercise. In each of the 27 Member States, the sample has been built by including banks, in descending order of size, so as to cover at least 50% of the respective national banking sector, as expressed in terms of total assets. As the stress test has been conducted on the highest level of consolidation for the bank in question, the exercise also covers subsidiaries and branches of these EU banks operating in other Member States and in countries outside Europe. As a result, for the remaining 7 Member States where more than 50% of the local market was already covered through the subsidiaries of EU banks participating in the exercise, no further bank was added to the sample. The 91 banks represent 65% of the total assets of the EU banking sector as a whole.

Given the differences in size and complexity, business models, scope of operations and risk profiles of the institutions included in the sample, it should be borne in mind that the aggregate results presented in this report cannot be directly applied to individual institutions, nor can be directly extrapolated to other banks in the EU. This point is of special importance as regards the assessment of banks’ continued reliance on government support measures, as the sample of banks contains both institutions making use of various support measures, and institutions which did not revert to public support.

Risk factors included in the stress test exercise

The stress test focuses mainly on credit and market risks, including the exposures to European sovereign debt. The focus of the stress test is on capital adequacy; liquidity risks were not directly stress tested.

The exercise has been carried out on the basis of the consolidated year-end 2009 figures and the scenarios have been applied over a period of two years - 2010 and 2011.

Scenarios used in the exercise

For the purpose of stress testing the credit risk and simulating the profit and losses, two sets of macro-economic scenarios (benchmark and adverse) have been developed, in close cooperation with the ECB and the EU Commission. The benchmark scenario was based on the EU Commission Autumn 2009 forecast and the European Commission Interim Forecast in February 2010, with several adaptations to reflect recent macro-economic developments in a number of countries. The adverse macro-economic scenario was based on ECB estimates. Within the adverse scenario, the exercise also envisages a “sovereign risk shock”, reflecting adverse conditions in financial markets.

For each macro-economic scenario, a set of key macro-economic variables (including GDP, unemployment, interest rate assumptions) was provided for the domestic situation for each EU Member State, the US, and the rest of the world collectively. Some of the input parameters and assumptions have been provided by CEBS, and by the participating supervisory authorities outside of the narrative of the macro-economic scenarios as provided by the EU Commission and the ECB, notably the evolution of the real estate prices.

The benchmark macro-economic scenario assumes a mild recovery from the severe downturn of 2008-2009, whereas the adverse scenario assumes a “double-dip” recession. For the euro area, the GDP growth under the benchmark scenario is assumed at a level of +0.7 (2010) and +1.5% (2011), whereas under the adverse scenario the euro area would see a decrease of GDP by -0.2% in 2010 and -0.6% in 2011. For the whole European Union (EU27) the benchmark scenario assumes a +1.0% growth of GDP in 2010 and +1.7% in 2011, whereas under the adverse scenario the GDP would not grow in 2010 and would decline by -0.4% in 2011. On aggregate and over the two-year time horizon, the adverse scenario assumes a three percentage point deviation of GDP for the EU compared to the benchmark scenario. It should be noted that current macro-economic developments remain in line with the assumptions provided in the benchmark scenario.
In addition to a global confidence shock, that affects demand worldwide, the adverse scenario envisages an EU-specific shock to the yield-curve, originating from a postulated aggravation of the sovereign debt crisis. The latter impact is differentiated across countries, taking into account their respective situation.

In particular, related to prevailing sovereign debt risks, a common upward shift in the yield curve was applied for each country in the EU (reaching 125 basis points for the three-month rates and 75 basis points for the 10-year rates at end-2011), supplemented with country-specific upward shocks to long-term government bond yields (overall amounting to 70 basis points at end-2011 for the euro area). The rise in short-term rates reflects an assumption of tensions in the interbank market – as was seen during earlier financial turmoil episodes. The country-specific bond yield shock in turn accounts for differentiated fiscal situations and related market perceptions. This results in a set of haircuts to be applied to all EU sovereign bond holdings in the trading books of the banks in the sample.

For the purposes of the market risk stress test, a set of stressed market parameters was applied to the trading book positions. It should be noted that the parameters developed for the market risk stress test are in-line with the macro-economic scenarios, and therefore could be considered as directional, meaning that depending upon the size and direction of their exposures, banks were able to make gains on certain portfolios, thereby reducing the overall amount of stress coming from the market parameters.

Key common assumptions used in the exercise

The exercise was conducted, using common templates, a common methodology and under key common assumptions. In particular, the exercise assumes, both for the benchmark and for the adverse scenarios, a “zero growth” assumption for the evolution of exposures for market and credit risks over the whole stress horizon. However, any regulatory imposed decisions (e.g. restructuring plans agreed with the EU Commission under the State Aid reviews) as well as management actions (e.g. capital raisings or divestment programmes) publicly announced before 1 July 2010 have also been taken into account. The results do not include any government support of recapitalisation measures taken after 1 July 2010.

In conducting the exercise, for the major cross-border banking groups, the macro-economic scenarios were translated using internal models, internal risks parameters and granular portfolio data, whereas for the less complex institutions more simplified approaches were used in general (e.g. use of the reference parameters provided by the ECB for instance).

Securitisation positions have been tested under the assumption of rigorous and uniform reductions in the credit quality of the positions as of end 2009, which already incorporated very material reductions in external credit ratings, as compared to their original level. For the adverse scenario, the assumed reduction in credit quality of the positions is equivalent to four external rating notches over two years. The impact of such reduction has been recorded as an increase in risk-weighted assets (the denominator of the solvency ratio) and as a direct reduction of regulatory capital (the numerator of the solvency ratio).

Equity exposures in available for sale portfolios have been subject to a cumulative haircut over two years of 19% in the benchmark scenario, and 36% in the adverse scenario. Other exposures in available for sale portfolios (i.e. bonds and loans) have been tested along with other credit exposures in the banking book.

In light of these assumptions, the information provided for the benchmark and forecast scenarios should in no way be construed as forecasts.

Although the exercise did not prescribe any specific restrictions to the profitability of operations and reduction of income, especially generated in the regions not covered directly by the macro-economic scenarios, the assumptions and forecasts used by the banks have been challenged by the respective national supervisory authorities and brought to the attention of CEBS.

Aggregate results

Based on the results of the calculations, the aggregate Tier 1 capital ratio, used as a common measure of banks’ resilience to shocks, would decrease under the adverse scenario including sovereign shock from 10.3% in 2009 to 9.2% by the end of 2011. It should be noted that the aggregate Tier 1 capital ratio incorporates approximately 169.6 bn € of government capital support provided until 1 July 2010, which represents approximately 1.2 percentage point of the aggregate Tier 1 capital ratio. It should be noted that the maturity of government support measures extended to banking institutions in the sample goes way beyond the two-year time horizon of the exercise. As such, government support form an integral and stable part of the Tier 1 capital ratios of the banks in question. It is not expected that any withdrawal of government support measures could take place without appropriate substitution by private funding sources, where relevant.

The downward pressure on capital ratios under the adverse scenario including sovereign shock is mostly stemming from impairment losses (472.8 bn € over the two-year period) and trading losses (25.9 bn € over the two-year horizon). Losses associated with the additional sovereign shock would reach 67.2 bn € over the two-year period (among which 38.9 bn € associated with valuation losses of sovereign exposures in the trading book). In total, aggregate impairment and trading losses under the adverse scenario including the additional sovereign shock would amount to 565.9 bn €.

The average two-year cumulative loss rates associated with these losses are 3.0% for corporate exposures and 1.5% for retail exposures under the benchmark scenario, and 4.4% for corporate and 2.1% for retail exposures under the adverse scenario, compared with average loss rates of 1.5% for corporate exposures and 0.8% for retail exposures in 2009.

As a result of the exercise, under the adverse scenario 7 banks would see their Tier 1 capital ratios fall below 6%, with an overall shortfall of 3.5 bn € of Tier 1 own funds. The threshold of 6% is used as a benchmark solely for the purpose of this stress test exercise. This threshold should by no means be interpreted as a regulatory minimum (according to the CRD2 the regulatory minimum for the Tier 1 capital ratio is set to 4%3), nor as a capital target reflecting the risk profile of the institutions, the latter being the outcome of the supervisory review process under Pillar 2 of the CRD.

The CRD regulatory minimum for the overall capital adequacy ratio is set to 8% with a minimum Tier 1 capital adequacy ratio set to 4%. Several EU Member States have opted for higher minimum capital adequacy ratios.

The aggregate results suggest a rather strong resilience for the EU banking system as a whole and may appear reassuring for the banks in the exercise, although it should be emphasized that this outcome is partly due to the continued reliance on government support for a number of institutions. However, given the uncertainties over the actual path of the macro-economic recovery, the result should not be seen as a reason for complacency.

The adverse macro-economic developments seen in 2008-2009 (EU27 GDP falling by -4.2% in 2009) led to record high loan losses reported in 2009, whereas in early 2010 we witnessed improved macro-economic conditions which suggest an increase of capital ratios attributed to higher retained earnings affected by lower loan losses. In addition, it should be noted that interest assumptions of the macro-economic scenarios, while having a minor impact on the loan losses, may have a sizable offsetting impact on the income side, leading to an increase of net interest income in some cases and thus positively affecting the profitability of some banks. Last, but not least, many of the banks in the exercise have significant operations outside the EU. Some of these countries have weathered the crisis comparably well and continue to show strong economic growth. Further, increased revenue streams from those economies positively contribute to these banks overall profitability, offsetting loan losses and building sizeable retained earnings.

Follow-up on the stress test results

Part of the mandate of CEBS is to undertake on a periodic basis these EU-wide stress testing exercises. CEBS will continue with testing the resilience of the EU banking sector by means of periodic EU wide and thematic risk assessments and stress testing exercises, and will continue its work on improving convergence in supervisory practices across Europe by addressing the topics both from a policy and practical perspective.

CEBS supports the greater transparency of this exercise and of the results of this stress test exercise, given the specific market circumstances under which banks currently operate and thus welcomes the decision to publish bank individual results, as well as detailed information on banks’ exposures to EU sovereign debt.

With respect to the situation of individual institutions that fail to meet the threshold for this stress test exercise, the competent national authorities are in close contact with the banks in question to assess the results of the test and their implications, in particular any potential need for recapitalisation.

The banks are expected to propose a plan to address the weaknesses that have been revealed by the stress test. The plan will have to be implemented within an agreed period of time, in agreement with the supervisory authority.
Details of the follow-up actions are provided at national level by the supervisory authorities.