Comprehensive Assessment Stefan Kerbl, Principal Expert, ECB Sunday 26 th October 2014

Rubric 1 The three primary aims of the Comprehensive Assessment are transparency, repair and confidence building Repair Confidence building Transparency Identify and implement necessary corrective actions, if and where needed Enhance the quality and quantity of information available on the condition of banks Reassure all stakeholders that SSM banks are fundamentally sound and trustworthy More broadly the exercise aims to facilitate banks’ provision of credit to the European economy by reducing uncertainty over their solvency Introduction to the comprehensive assessment

Rubric 2 The Comprehensive Assessment comprises of two main components – an Asset Quality Review and a Stress Test 1. Asset Quality Review In depth review of banks’ balance sheets: ‒ Assessment of data quality, asset valuations, classifications of non-performing exposures, collateral valuation and provisioning ‒ Covers credit and market exposures, following a risk-based, targeted approach 2. Joint ECB / EBA Stress Test Forward-looking view of banks’ shock-absorption capacity under defined baseline and adverse scenarios Conducted in collaboration with the European Banking Authority (EBA) Results of the Asset Quality Review were incorporated into the Stress Test Introduction to the comprehensive assessment

Rubric 3 Comprehensive in scope Comprehensive in nature 19 participating countries 130 participating banks 82% of total SSM banking assets covered Every portfolio of every participating bank examined and a risk-based sub-set selected for review in extensive detail 816 individual portfolios examined in detail 119,000 debtors analysed in detail 170,000 collateral items (re)valued 825 provisioning models challenged 4,500 securities revalued 70 derivative pricing models reviewed The exercise was comprehensive in its scope and nature Granular analysis with 40 million data points for the Stress Test alone, equivalent to 300,000 per bank Introduction to the comprehensive assessment

Rubric 4 Scenario building blocks: four main sources of risk starting with global vulnerabilities Global sources of risk Increased risk aversion, broad-based sell-offs and re-pricing EMEs specifically impacted, with a severe decline in world trade Currency depreciation and funding stress in CEEs Domestic demand confidence-driven shocks Sovereign bond yield differentiation Access to and costs of bank funding impacted Stress Test recap

Rubric a comprehensive methodology prescriptive also for parts of the stress that are less focussed on previously (NII, PPP) 2. a systematic and centrally-led quality assurance process Thoroughly defined QA approach in the CAST manual Each bank had to provide up to 307,000 data points (circa 40 million in aggregate) Involving 100FTEs at central ECB level and hundreds of FTEs across the SSM 3. The thorough assessment of starting values via the AQR a join-up process for the AQR and stress test outcomes Both book values and risk parameters Mechanistic and comprehensive Detailed description of the methodology in the CAST manual Three important features distinguish the ECB CA stress test from previous EU-wide stress test exercises Stress Test recap

Annex

Rubric 7 The join-up process connects and reinforces the ‘point-in-time’ AQR and the ‘forward-looking’ stress test, thereby strengthening the overall exercise The join-up is in line with some previous national exercises (e.g. Spain, Cyprus, Slovenia) – but unparalleled on the euro area level The AQR findings impact on the starting point CET1 ratio of all banks Where the AQR identifies material issues, these is also used to adjust bank’s forward looking stress test results (e.g. loan losses) AQR results will be used to adjust starting point and stress test projections Stress Test recap

Rubric Quality assurance (QA) was a key ingredient for ensuring credible results from a constrained bottom-up stress test Enhance accuracy and credibility by providing a systematic review of individual bank results Key objective to minimise the risk that banks passing the stress-test will subsequently fail Also key for ensuring a level playing-field in a multi-country context Quality Assurance scope: Stress test results (baseline and adverse scenario) Stress test results including the join-up impact Efficient framework for QA by focusing on material issues (traffic light approach) Stress Test recap

Rubric 9 A central QA framework has been put in place to review results from individual and cross-country perspectives and by risk dimension The central QA framework Specify a number of quantitative checks to be performed on banks’ submitted results. These constitute a minimum standard to be applied consistently across all countries, with Red/ Amber/ Green (RAG) thresholds. employs ECB top-down stress test models, in-market and cross-market comparisons harnesses complementary qualitative information on each bank (e.g. reviewing explanatory notes) Three lines of defence: NCA bank teams – NCA centrally – ECB ECB CONFIDENTIAL Stress Test recap

On the Results

Rubric Comprehensive Assessment results

Rubric 12 Comprehensive Assessment - key figures Key results The Asset Quality Review (AQR) results in a gross impact on asset carrying values of €48 billion In total, a €136 billion increase in non-performing exposure was identified Combining the AQR with the stress test the Comprehensive Assessment results in: - €263 billion capital depletion over the three-year horizon of the exercise under the adverse stress test scenario -Median 4% reduction of the CET1 capital ratio of in scope banks In aggregate, the Comprehensive Assessment resulted in a €24.6 billion capital shortfall across 25 participant banks Comprehensive Assessment results

Rubric Comprehensive Assessment results Note: Numbers do not add up due to rounding Comprehensive assessment identified a capital shortfall of €24.6 billion across 25 banks +10.7BN Comprehensive assessment capital shortfall by driver SSM level (€ BN) +2.7BN 13

Rubric Capital shortfall was observed at banks from 11 of the 19 countries in scope of the exercise Comprehensive assessment capital shortfall by driver By country, as % RWAs Total shortfall (€ BN) Comprehensive Assessment results

Rubric Comprehensive Assessment results Median bank’s CET1 ratio declines from 12.4% to 8.3% Comprehensive assessment impact on CET1 ratio under the adverse scenario Median by country of participating bank, % SSM median: 4.0% The median bank’s CET1 ratio falls by 4% in the adverse scenario 15

Rubric Asset Quality Review results

Rubric 17 Across the SSM, the Asset Quality Review (AQR) led to a €48BN adjustment to asset carrying values Asset Quality Review impact on available CET1 capital By AQR workblock (€ billion) Additional provisions Other capital adjustments Projection of findings Impact from risk-based sample Asset Quality Review results

Rubric 18 The AQR led to an €136 BN increase in non-performing exposure, with increases across all asset segments Change in NPE exposure, pre- and post-AQR By asset segment (€ billion) Divergent bank definitions of non- performing exposures were harmonised leading to €55 billion added non- performing exposure Following harmonisation, an increase in non- performing exposure of €81 billion was observed in the credit file review In total, non-performing exposure increased by €136 billion, representing a 18% total adjustment Commentary NPE exposure (€BN) Individually assessed (credit file review) Collectively assessed (collective provisioning) +33% +19% +18% +31% +36% +73% +14% +4% +1% Asset Quality Review results

Rubric 19 Individually assessed (credit file review) Collectively assessed (collective provisioning) Provisioning increased by a total €43 BN across all asset segments Change in provisions By Asset Segment (€ billion) Total specific provisions increased by €43 billion, a 12% overall adjustment Provisions increased as a result of both the credit file review and collective provisioning workblocks Shipping (28%), Large SME (16%) and Large Corporates (16%) experienced largest relative increases Commentary Provisions (€BN) +16% +10% +28% +10%+5% +12%+6%+5% Asset Quality Review results

Rubric ECB Quality Assurance had a tangible impact on NPE classification, ensuring harmonised treatment ECB identified banks in where debtors were hitting triggers but not being classified as NPE ECB discussed with NCAs and challenged auditor decisions at the individual debtor level In some cases the decision against reclassification was justified In a significant number of cases, decision was withdrawn and the debtor reclassified to NPE along with debtors in similar scenarios Example of impact of ECB Quality Assurance Number of performing debtors hitting 2 or more impairment triggers, pre- and post- ECB Quality Assurance (%) Remedial approach taken 20 Asset Quality Review results

Rubric ECB Quality Assurance resulted in a significant increase in collateral haircut levels Example of impact of ECB Quality Assurance Mean collateral haircuts pre- and post- ECB Quality Assurance (%) Note: The exhibited number of banks is not necessarily exhaustive for the example NCA ECB reviewed haircut levels across NCAs for each asset segment ECB discussed with NCAs and challenged auditor decisions at the individual debtor level In some cases the ECB accepted the NCA submission In others additional haircuts were agreed and applied Remedial approach taken Asset Quality Review results 21

Rubric In total, collective provisioning led to an increase in provisions of €16BN, of which 62% was IBNR Collective provisioning adjustment – IBNR SSM-level, € billion Collective provisioning adjustment – specific provisions SSM-level, € billion In total, more than 800 portfolios across most AQR asset classes were assessed Collective Provisioning workblock identified the need for additional collective provisions of €16 billion, - €6 billion of retail specific provisions - €10 billion of additional IBNR Key drivers included Application of EBA simplified NPE definition Credit file review findings leading to adjustments in LGI parameter Adjustments to RRE collateral values impacting LGL Bank use of non point-in-time parameters +23% +6% Asset Quality Review results 22

Rubric Collective provisioning Quality Assurance aligned parameters to ECB defined fall back assumptions Collective provisioning parameter distribution – emergence period Distribution of performing exposures by emergence period Asset Quality Review results ECB defined fall back assumption ParameterFall back assumption Observed average LGL secured 60%50.4% LGL unsecured 90%86.9% Original effective interest rate 4%3.6% Sales ratio 75%78.0% Sales ratio volatility 18%21.6% Appraiser discount 5%5.4% Comparison of other fall back parameters Number of exposures 23

Rubric The adjustment of the Fair value exposures review was €4.6 billion, with 66% from CVA adjustments Fair value exposures review adjustment By workblock (€ billion) Non-derivative positions were assessed through independent revaluations leading to a €1.2 billion adjustment Adjustment on CVA reserves was significant, with a 27% increase of €3.1 billion identified Complex derivative pricing models were also reviewed, with modelling errors or inappropriate assumptions leading to a further €0.2 billion adjustment Asset Quality Review results 24

Rubric Stress Test & Join-up results

Rubric 1 Stress Test results include the impact of the Join-Up Note: Scenario capital depletion and the effect on required capital are based on the 2016 adverse scenario Overall, total adverse scenario capital depletion is €263 billion Contribution of the Stress Test Comprehensive assessment adverse scenario capital depletion SSM level, (€ BN) Key drivers Total gross AQR adjustment of €48 billion, and €34 billion net of tax offset The stress test (and Join-up with AQR results) led to a capital depletion of €182 billion in the adverse scenario In addition, the increase in RWA in the adverse scenario increases capital requirements in the amount of €47 billion Gross AQR adjustment 26

Rubric SSM banks' average CET1 ratio is projected to increase from 11.8% to 12.0% in the baseline Key drivers Improvement in the solvency position under the baseline mainly reflects –Projected accumulation of pre-provision profits (3.6 percentage point contribution to the change in the CET1 ratio) –Projected loan losses (-2.5 percentage point contribution) The average development of participating banks’ solvency positions, however, masks variations across individual institutions and countries Stress Test results Aggregate post-JU stress test effect 1 by risk drivers under the baseline scenario 1. Weighted means; excluding the AQR impact on starting point capital 27

Rubric SSM banks' average CET1 ratio is projected to decrease from 11.8% to 8.8% in the adverse Key drivers Increase in loan losses (-4.5 percentage point contribution to the change in the CET1 ratio) Lower pre-provision profits compared to the baseline (corresponding to a 1.3 percentage point lower positive contribution the change in the CET1 ratio) “Administrative and other expenses” have an impact on the overall results; however, they remain largely unchanged between the baseline and adverse scenario and mainly reflect staff and other administrative costs that regardless of the scenario have a negative impact on banks' loss absorption capacity Stress Test results Aggregate post-JU stress test effect by risk drivers under the adverse scenario 1. Weighted means; excluding the AQR impact on starting point capital 28

Rubric Loan losses and net interest income are key drivers of divergence from baseline to adverse Stress Test results Aggregate post-JU stress test effect by risk drivers under the adverse scenario 1. Weighted means; excluding the AQR impact on starting point capital 29

Rubric Corporate and retail portfolios are the key drivers of loan losses in both scenarios Key drivers Loan losses across banks are mainly driven by the corporate and retail portfolios, both under the baseline and adverse scenarios Under the baseline scenario, the median CET1 percentage point reduction due to losses is: –0.9% in the corporate segment –0.5% in the retail segment Results under the adverse scenario are, however, more severe with a median CET1 percentage point reduction of –1.6% in the corporate segment –1.1% in the retail segment Stress Test results Baseline scenarioAdverse scenario Decomposition of loan losses across portfolios and banks under the baseline and adverse scenario 30

Rubric Under the adverse scenario, the median decline in NII is larger and more varied across banks Key drivers While the picture is heterogeneous across banks, the median decline in net interest income is larger under the adverse than the baseline scenario Moreover, the distribution of changes in net interest income across banks is in general wider under the adverse scenario Stress Test results Net interest income development across banks under the baseline and adverse scenario, year-on-year % changes 31

Rubric RWAs grow in net terms across the horizon, resulting in higher capital requirements Stress Test results RWA development across banks under the baseline and adverse scenario, year-on-year % changes Key drivers Risk-weighted assets experience net growth across the horizon, albeit at a declining rate For the large majority of banks under the static balance sheet assumption, the nominal balance sheet size remains the same by design Risk weights for the median bank grow under the baseline scenario from 1.0% in the first year to 0.7% in the third year, and under the adverse scenario 3.2% in the first year to 0.9% in the third year Increased RWAs result in higher capital requirements 32

Rubric The stress test impact differs across banks under the static and dynamic balance sheet assumption Stress Test results Distribution of changes to CET1 ratios across banks following the static vs. dynamic balance sheet assumption under the baseline and adverse scenario, cumulative % changes Key drivers Banks under the dynamic balance sheet assumption are less heavily affected under the baseline scenario In the adverse scenario larger CET1 ratio declines are observed for banks under the dynamic balance sheet assumption. This could reflect that restructuring banks –Are generally weaker and more vulnerable to stress tests –May be located in countries with relatively more severe scenarios In cases where banks provided both, static and dynamic templates, the dynamic version generally resulted in less severe effects 33

Rubric Join up effect varies by bank but is driven by bank AQR impact Join-up effect by bank in relation to AQR impactKey drivers Join up effect is highly correlated with the magnitude of AQR findings The strongest join-up effect (above 1% of RWA) is observed for banks where AQR had a major impact For banks with small or negligible AQR findings, the join-up effects on average were similarly small (<0.2% of RWA) Join-up results 34

Rubric Impairments are the major driver of join-up effect by change in CET1 in the baseline scenario Join-up results CET1 effect of join-up by type (credit vs. other effects) under the baseline scenario 35

Rubric Distribution of join-up effect by type is similar, but for greater impacts overall, in the adverse scenario Join-up results CET1 effect of join-up by type (credit vs. other effects) under the adverse scenario 36

Rubric The post-JU impact of the Stress Test is 0.2% in the baseline and -3.0% in the adverse Stress test component (€ billion) Stress test results (post-JU) BaselineAdverse NII Net fee and commission income Net trading income 256 Sovereign FVO/AFS -28 Admin. and other expenses -865 Loan losses Taxes, dividends and other Total CET1 impact (€ billion) Total CET1 ratio change (percentage points) 0.2%-3.0% of which: Join-up CET1 impact (€ billion) Stress Test and Join-up results 37

On the Scenario

Rubric 39 PROCESS / STORY DERIVED RESULTS 1 4 ASSUMPTIONS / DRIVERS2 RESULTS / SEVERITY3

Rubric 40 Scenario building process The baseline scenario was prepared by the European Commission The adverse scenario was proposed by the ESRB, working in close collaboration with the ECB and the EBA, and it was finally approved by the EBA Board of Supervisors It captures the prevailing view of systemic risks facing the EU financial system, as identified by the ESRB General Board It includes forward-looking paths for key macroeconomic and financial variables for all EU countries and a large number of non-EU countries over a three-year horizon

Rubric Starting point: global sources of risk -Increased risk aversion, broad-based sell-offs and re-pricing -EMEs specifically impacted, with a severe decline in world trade -Currency depreciation and funding stress in CEEs Triggering EU-specific risks -Domestic demand confidence-driven shocks (real estate too) -Sovereign bond yield differentiation re-appearing -Access to and costs of bank funding impacted Scenario narrative 41

Rubric 42 Scenario building blocks: Four main sources of risk leading to a set of drivers Source of risk:Financial and economic shocks: Increase in global bond yields amplified by an abrupt reversal in risk assessment, including towards EMEs, and pockets of market liquidity  Financial market shocks worldwide (sovereign bonds, corporate bonds, stock prices, etc.).  Demand shocks in EMEs  EU countries: foreign demand shocks via a decline in world trade  Currency depreciation and funding stress in CEEs Further deterioration of credit quality in countries with feeble demand, with weak fundamentals and still vulnerable banking sectors  EU country-specific aggregate demand shocks (via fixed capital formation and private consumption)  EU country-specific aggregate supply shocks (via shock on user cost of capital, nominal wages)  EU country-specific house price shocks Stalling policy reforms jeopardising confidence in the sustainability of public finances  EU country specific sovereign bond spread shocks Lack of necessary bank balance sheet repair to maintain affordable market funding  EU-wide shock to short-term interbank interest rates  EU country-specific shocks to borrowing costs for households and corporates (via shocks to respectively, wealth and user cost of capital)

Rubric 43 PROCESS / STORY DERIVED RESULTS 1 4 ASSUMPTIONS / DRIVERS2 RESULTS / SEVERITY3

Rubric 44 A selection of shocks – aggregated summary Shocks (deviations from baseline, in basis points or percentage)

Rubric 45 Equity price shocks for EU-countries – country specific (percentage deviation from baseline levels in 2016)

Rubric 46 Adverse scenario – respective role of ‘drivers’ Global shocks account for c. 40% of the total GDP impact Shocks to domestic real economic variables account for c % of the total GDP impact Sovereign debt shocks and bank funding shocks together account for c % of the total GDP impact Contributions of scenario components to GDP impact (percentage point deviation from baseline levels in 2016)

Rubric 47 PROCESS / STORY DERIVED RESULTS 1 4 ASSUMPTIONS / DRIVERS2 RESULTS / SEVERITY3

Rubric 48 Key scenario variables – output for the areas Deviations from baseline levels by end-2016

Rubric 49 Cumulative GDP impacts in previous CEBS/EBA stress test exercises (Deviation of adverse from baseline growth rates in percentage points) Comparison with past EU-wide stress tests

Rubric 50 Cumulated impact on domestic GDP (percentage deviation from baseline) Domestic GDP levels incl. baseline (pre-stress Q0 GDP is scaled to 100) Pre-stress GDP level Assessing the severity of the scenario w.r.t. CCAR Domestic GDP scenarios

Rubric 51 PROCESS / STORY DERIVED RESULTS 1 4 ASSUMPTIONS / DRIVERS2 RESULTS / SEVERITY3

Rubric 52 Market risk benchmarks – derived from the scenario Derived as a function of and consistent with the underlying macro-financial baseline and adverse scenarios; no additional exogenous shocks 170 market risk parameters including interest rates, equity prices, commodity prices, swap spreads, credit spreads, dividends, as well as volatilities of market risk parameters For major economies such as the euro area, the United States, the United Kingdom, Japan, other developed markets, and major emerging market regions Next to historical “scenarios”; different exercise

Rubric 53 Market risk benchmarks – already available Calibrated market risk parameters

Rubric 54 Credit risk benchmarks – to come next… Derived as a function of and consistent with the underlying macro-financial baseline and adverse scenarios; no additional exogenous shocks Loss rates projections computed for the EU countries and 20 non-EU countries and regions Main portfolio segments includes the non-financial corporate sector (real-estate related and non- real-estate related), household mortgage loans, consumer credit, as well as financial institutions.