## chwebfea ## Source details **Canonical URL:** [chwebfea](https://www.imf.org/-/media/files/publications/dp/2021/english/chwebfea.pdf) ## Other formats - [Markdown version](/-/media/files/publications/dp/2021/english/chwebfea.pdf.md) - [Structured JSON version](/-/media/files/publications/dp/2021/english/chwebfea.pdf.json) --- ### Executive Summary — Overall impact and baseline - COVID-19 exacts a severe social and economic toll on Europe; banks face exposure to hard-hit sectors and pre-existing profitability weaknesses. - The paper evaluates impact on bank capital under a range of macroeconomic scenarios using granular sectoral exposures and incorporating pandemic-related policy support. - Coverage: wider range of policies and broader set of banks—including smaller banks—within the euro area and most European countries outside the euro area. Key baseline findings and scenario results - Aggregate CET1 outcomes: - Aggregate CET1 ratio (end-2019): 14.7 - Projected CET1 ratio (end-2021, baseline with policy measures): 13.1 - Projected CET1 ratio (end-2021, baseline without policy measures): 11.5 - IMF exercise: aggregate capital impact over two years of 1.7 percentage points (with policy measures) - Resilience and risks: - No aggregate capital shortfall relative to minimum prudential requirement, but some large euro area banks may struggle to meet MDA thresholds, creating funding pressures for hybrid capital. - Cross-country variation is considerable; capital change sensitive to shock size and initial balance-sheet condition. - Policy materially reduces extent and variability of capital erosion; without policies capital erosion would be 1.7 percentage points larger. - Adverse scenario: - Adverse scenario with slower recovery in 2021 yields more pronounced capital erosion and potential doubling of banks breaching MDA thresholds. - Aggregate additional CET1 decline in adverse scenario: 1 percentage point by end-2021 even with current policy measures. ### Bank vulnerabilities and profitability - Pre-existing conditions: - CET1 capital: nearly 15 percent of risk-weighted assets in 2019. - Non-risk weighted leverage ratio: around 6 percent average in 2019; minimum threshold of 3 percent binding from June 2021. - NPLs: peaked at 7½ percent after GFC in 2013; declined to about 3 percent by 2019. - Profitability: ROA for many European banks declined since GFC; euro-area banks’ ROA below peers in other advanced economies in 2019. - Pandemic exacerbation: - Banks will likely raise provisions, write off rising NPLs, and face lower non-lending income. - More than 60 percent of banks’ corporate exposures are to sectors highly affected by the pandemic. - Bank lending to firms in highly affected sectors ≈ 200 percent of Tier 1 capital on average; exceeds 250 percent in some countries. - More than half of bank lending is to households (mortgages), vulnerable to income and employment shocks. - Time to rebuild capital via profits after end-2021 (modelled to restore CET1 = 14.7%): - Precrisis RoA = 0.38 percent → 2.6 years - Half precrisis RoA = 0.19 percent → 5.2 years - Required RoA = 1.06 percent → 1 year - Implication: absent material rise in profitability, organic replenishment can take multiple years. ### Role and treatment of policy measures - Policy types modeled: debt repayment relief (moratoria), public credit guarantees, deferred insolvency proceedings, capital relief/conservation measures, dividend restrictions (only in 2020), and other borrower support (grants, wage subsidies, tax deferrals). - Policy effects: - With policies, aggregate CET1 declines are substantially smaller and dispersion across banks reduced. - Debt repayment moratoria, credit guarantees, and insolvency moratoria materially affect projected capital outcomes and cross-country dispersion. - Example policy magnitudes and usage: - 39 of 44 surveyed European countries offer pandemic-related guarantee programs. - 38 European countries introduced debt service moratoria. - Seven countries offer a total guarantee envelope well in excess of 10 percent of GDP. - Typical guarantee coverage ratio: 70 to 90 percent. - EU State Aid rule ceiling for 100 percent guarantees: €800,000 per company (increased to €1.8 million per company on January 28, 2021). --- ### Introduction — objectives, coverage, and methodology - Objective: quantify impact of pandemic and policy support on bank capital using IMF January 2021 WEO Update baseline and illustrative adverse paths; assess counterfactual without support measures. - Coverage: aim to reach at least 80 percent of domestic banking sector assets in euro area; include banks in non-euro area European countries. - Methodology: - Top-down stress testing using granular bank-specific corporate exposures mapped to sector- and country-level projected loss rates. - Three channels modelled: profitability, asset (write-offs and lending), and risk-weight (credit risk weights via IRB formula). - Satellite ROA model sensitivities (largest euro area banks): every 1 percentage point decline in real GDP growth → ROA −0.27 percentage point; every 1 percentage point unemployment increase → ROA −0.21 percentage point; 1 percentage point increase in lagged NPL ratio → ROA −0.0155 percentage point. - One-year PDs extracted from observed risk-weights using IRB mapping and used to shock RWs under stress. --- ### Euro Area: large banks’ exposures and regulator responses - Regulators’ relief and timeline: - Banks allowed to use combined capital buffers, temporarily operate below Pillar 2 guidance and liquidity coverage ratio; flexibility in classification and provisioning for loans backed by public support. - ECB supervisory flexibility introduced March 20; EBA guidance on default/forbearance and IFRS 9 application issued March 25 and April 2, 2020. - Labor and direct business support: - Short-time work schemes supplement up to 70–80 percent of employees’ pay for hours not worked in several economies. - Cash grants, tax deferrals, expanded direct lending provided in several countries. - Lending dynamics: - Euro area bank lending growth: 5 percent (yoy) at start of 2020 → nearly 9 percent (yoy) in May 2020. - New lending growth to firms ≈ 6½ percent (yoy) by October 2020. - Government guarantees lower capital costs via lower risk-weights. Debt service and insolvency moratoria - Debt service moratoria: - Introduced in 38 European countries; duration initially 3–6 months and extended in some countries into 2021. - Moratoria reduce immediate cash receipts and can entail economic losses beyond accounting treatment. - EBA guidance encouraged “look through” approach; temporary lapsed in mid-September and reinstated in December with emphasis on timely credit-risk recognition. - Insolvency moratoria: - Most countries adopted insolvency moratoria for about half a year from end-March 2020; cliff-edge risks when moratoria expire. - EA median duration of insolvency moratoria: 5.7 months; Other EU/EEA median: 0 months. --- ### Stress-test scenarios, channels, and quantitative results - Scenarios calibrated to January 2021 WEO Update: - Baseline: sharp 2020 recession followed by partial recovery in 2021. - Illustrative adverse: slower 2021 recovery if vaccinations slower and containment stricter/longer. - Example ECB projection: euro area GDP fell by 6.9 percent in 2020. - Channel decomposition (baseline, with policies): - Profitability channel: −0.7 percentage point CET1 - Asset channel: −0.5 percentage point CET1 - Risk-weight channel: −0.5 percentage point CET1 - Baseline and sensitivity outcomes: - Aggregate CET1 capital ratio for large euro area banks declines from 14.7 to 13.0 percent by end-2021 (baseline, accounting for granular borrower support measures) in one reported panel. - Capital erosion decreases by about one-third if banks refrain from dividend distributions in 2021. - Without supportive policies, capital erosion is roughly doubled. - About 14 percent of the largest (90) euro area banks likely to breach MDA (baseline, without policy support) in one estimate; with policies this share is markedly lower. - Potential capital needs: - If ~14 percent of large banks breach MDA without policies → required capital ~€25 billion (or 1.7 percent of reported CET1 at end-2019). - Without policies under adverse scenario → capital need nearly €47 billion (or 3.1 percent of reported CET1 at end-2019). Expanded sample (468 banks) - End-2019 CET1: - Expanded-sample euro area banks: 14.9 - EBA-covered large banks: 14.7 - Under baseline: - About 9 percent of all euro area banks in expanded sample likely to breach MDA if policies do not operate; required new capital: €26 billion (1.7 percent of reported CET1 at end-2019). - With supportive policies, about 3 percent of all euro area banks would struggle to clear MDA, capital shortfall €0.6 billion. - Non-EU CESEE banks: - End-2019 CET1: 12.8 - Projected CET1 (baseline): 10.8; (adverse): 10.5 - Mitigating policies lift CET1 by 0.4 percentage point in non-EU CESEE vs. ≈1.3 percentage points for EBA-covered euro area banks. Notable quantified NPL projection - Under baseline, NPLs of euro area banks could increase to more than €900 billion by 2021 (up from about €500 billion in 2019), raising the NPL ratio to more than 6 percent. --- ### Policy recommendations — multi-pronged strategy - Keep borrower support measures in place until recovery is firmly established; tighten eligibility as recovery gains momentum to target illiquid but viable firms and vulnerable households and prevent loan misclassification and “zombie” firms. - Clarify supervisory guidance on availability and duration of capital relief and conservation measures: - Allow banks to build back capital buffers gradually to avoid abrupt lending cuts. - Maintain restrictions on dividend payouts and share buybacks until recovery is well underway. - Support balance-sheet repair and resolution frameworks: - Strengthen NPL management and foster development of secondary markets for distressed assets; consider carefully designed AMCs where appropriate. - Strengthen insolvency regimes with fast-track debt restructuring and out-of-court workouts. - Use the current system-wide stress test (expected to be completed in July 2021) to assess precautionary recapitalization needs; precautionary recapitalizations could use Temporary State Aid Framework flexibility. - Address structurally low profitability: - Enhance non-interest revenues and improve cost structures. - Invest in digital technologies to streamline operations (recognize short-term expense increases). - Pursue domestic and cross-border consolidation to improve efficiency and allocation of capital and liquidity. Selected regulatory numeric thresholds and guidance - Regulatory minimum CET1 capital ratio: 4.5% - Capital Conservation Buffer (CCB): 2.5% (implied hurdle with regulatory minimum = 7.0%) - Average MDA threshold: about 9.1% of risk-weighted assets (example MDA labels: 9.1% and 10.6%) - Supervisory guidance: buffer use can be used through end-2021; Pillar 2 Guidance add-on replenishment not required until after 2022. - Stress-test milestone: system‑wide stress test expected to be completed in July 2021. --- ### Annex / Methodology — modelling details and parametrics - Three channels explicitly modelled: profitability (ROA decomposition), asset (net lending, write-offs), and risk-weight (IRB-based RW changes). - ROA satellite and NPL regression sensitivities (selected exact coefficients preserved): - Real GDP growth → change in ROA: 0.270*** (Model 1 SSM); 0.0752** (Model 6 All Banks). - Unemployment rate → change in ROA: −0.208** (Model 1); −0.0557 (Model 6). - NPL ratio (lagged) → change in ROA: −0.0155 (Model 1); −0.0295* (Model 6). - NPL ratio model (All Banks): real GDP growth coefficient 0.698***; NPL lag coefficient 0.431***. - IRB mapping and risk-weight calibration: - Capital requirement K uses asymptotic single risk factor IRB formula with statistical confidence a = 0.999. - Assumed average maturities: M_C = 6.3 [years]; M_H = 22.0 [years]. - Pre-stress provisioning rule-of-thumb: LLP = (1.662 + 0.00092 (RW×100)^2 − 0.06 (RW×100)) × LGD. - Loan growth and asset projection: - Credit supply equation includes lagged CET1 and lagged NPL ratio; predicted loan growth applied only if CET1 ≥ MDA threshold (0.091 for euro area average; 0.106 for non-euro). - Predicted reduction in lending: example aggregate impact cited of about 1.6 percentage points lower lending growth under baseline; about 3 percentage points if policies do not operate as expected. - Timing and staging: - On average, only 15 percent of the calculated corporate shock occurs in 2020 with remainder in 2021 due to deferred insolvency proceedings. - ν_t = min(9_m ; 1) used to split write-offs timing within two-year horizon based on insolvency stay duration. - Sample and data sources: - EBA Transparency Exercise (117 EU/EEA banks; focus on 90 euro area banks) and expanded coverage to 467/468 banks via S&P Global Market Intelligence and FitchConnect. - Macroeconomic projections from IMF World Economic Outlook; sectoral loss rates from Chapter 3 of October 2020 REO: Europe. *Source: Introduction (chapter), chwebfea - Introduction, IMF staff paper.* ### Executive Summary ������������������������������������������������������������������������������������������������������ ### Executive Summary ### Overall impact of COVID-19 on European banks - The COVID-19 pandemic is exacting a severe social and economic toll on Europe. - European banks have substantially raised their capital buffers over the years, but many suffer from chronically low profitability due to inefficient cost structures, compressed net interest margins, and the drag of legacy assets from the Global Financial Crisis (GFC) and the European sovereign debt crisis. - Banks stand heavily exposed to economic sectors that have been hard hit by the pandemic. - The paper evaluates the impact of the crisis on European banks’ capital under a range of macroeconomic scenarios, using granular data on the size and riskiness of sectoral exposures and incorporating pandemic-related policy support. - Compared to previous studies conducted by the European Central Bank (ECB) and European Banking Authority (EBA) in 2020, the analysis covers a wider range of policies and a broader set of banks—including smaller banks—within the euro area, while extending the sample to cover most European countries outside the euro area. ### Key baseline findings and scenario results - The baseline results suggest that despite a significant fall in capital ratios, banks remain broadly resilient to the shock. - There is no aggregate capital shortfall relative to the minimum prudential requirement. - A number of the larger euro area banks may struggle to meet their threshold for the maximum distributable amount (MDA), which could create funding pressures, especially with respect to hybrid capital. - The data reveal considerable cross-country variation; the change in bank capital is sensitive both to the size of the macroeconomic shock and the initial condition of bank balance sheets and profitability. - Policy is extremely important in reducing both the extent and variability of capital erosion; in particular, good policies can substantially weaken the link between the macroeconomic shock and bank capital. - In an adverse scenario with a slower recovery in 2021, the erosion of bank capital would become more pronounced, especially if a premature phase-out of support measures increases default risk. - The number of banks potentially breaching their MDA threshold would double, greatly increasing the risk of higher capital costs and funding difficulties. - The paper notes the EU authorities should use the current system-wide stress test, expected to be completed in July 2021, to assess the need for precautionary recapitalizations. - Illustration: time to restore precrisis capitalization (CET1 = 14.7%) through profits after end-2021 is modeled (see referenced figure). ### Bank vulnerabilities and profitability challenges - Banks are likely to face rising capital and liquidity pressures due to shrinking profits and deteriorating asset quality. - The crisis has intensified pre-existing profitability challenges: banks are likely to raise provisions for higher loan losses due to lower average borrower quality; write off a rising share of NPLs to insolvent borrowers with diminishing prospects for collateral recovery; and face lower income from non-lending activities. - Balance sheet pressures could hinder banks’ ability to support credit growth for the recovery. - European non-financial firms exhibit greater dependence on loans than companies in other advanced economies, so constraints to bank credit supply would create correspondingly larger challenges for the economic recovery. ### Role and treatment of policy measures in the analysis - The analysis incorporates the important role of pandemic-related policy support, including regulatory relief for banks and policies to support businesses and households, which act to shield the financial sector from the real economic shock. - Various national and European-level policy measures adopted in 2020 have helped cushion the adverse economic impact of the crisis. - Policy measures considered include borrower support measures (e.g., debt repayment relief or “moratoria”, credit guarantees, direct support for firms), capital relief and conservation measures, and insolvency moratoria. - Policy assumptions materially affect projected capital outcomes and cross-country dispersion of results. ### Policy recommendations (multi-pronged strategy) - Keep in place borrower support measures, such as debt repayment relief or “moratoria”, credit guarantees, and direct support for firms, until the recovery is firmly established. - As the recovery gains momentum, eligibility criteria should be tightened to better target illiquid but viable firms and the most vulnerable households while preventing loan misclassification, credit misallocation and the rise of “zombie” firms. - Clarify supervisory guidance on the availability and duration of capital relief and conservation measures. - Banks should be allowed to build back capital buffers gradually, so new lending does not need to be cut back. - Restrictions on dividend payouts and share buybacks should be maintained until the recovery is well underway. - Support balance sheet repair by strengthening non-performing loan (NPL) management and the bank resolution framework. - As policy measures such as insolvency moratoria expire, a wave of bankruptcies and loan defaults are likely to follow. - The EU authorities should use the current system-wide stress test, expected to be completed in July 2021, to assess the need for precautionary recapitalizations. - Insolvency regimes should be strengthened, focusing particularly on fast-track procedures to restructure debt. - Address structurally low bank profitability. - Rising impairments and provisions will exacerbate the pre-existing challenge of very low rates of return on assets, limiting the ability of banks to restore capital buffers organically. - Banks should enhance non-interest revenues and improve their cost structures. - Investments in digital technologies to streamline operations are appropriate but may increase expenses over the short term. - Further domestic and cross-border consolidation could improve banks’ efficiency, while also facilitating a better allocation of capital and liquidity within banking groups. *Executive Summary — COVID-19: HOW WILL EUROPEAN BANKS FARE?* ### Introduction ### Introduction ### Context and objective - European governments introduced a range of exceptional mitigation measures to alleviate liquidity stresses of firms and households; the ECB-Banking Supervision and EBA announced capital relief and conservation measures (supplemented by the reduction of macroprudential capital buffers by competent national authorities). - Thanks to these measures, to date, the impact on bank capital has been relatively limited; corporate credit growth increased substantially during the initial phase of the pandemic, especially to highly affected sectors with higher liquidity needs. - This paper aims to quantify the impact of both the pandemic and policy support on bank capital and investigates the implications of the baseline and adverse GDP growth and unemployment paths published in the IMF’s January 2021 World Economic Outlook (WEO) Update on bank capital at the end of 2021. - The paper also assesses a counterfactual: what would have been the bank capital level without support measures such as repayment moratoria, loan guarantees, insolvency moratoria, and borrower measures? - Coverage objective: widen analysis beyond larger banks to reach a coverage of at least 80 percent of domestic banking sector assets in the euro area and include banks in some non-euro area European countries not covered by the EBA sample. ### Methodology and scope - Builds on previous analyses by European authorities (EBA 2020e; ECB 2020a) with a wider sample. - Incorporates sector-specific shocks to the real economy and a more comprehensive range of country-specific borrower support measures. - Uses granular data on bank-specific corporate exposures mapped to projected loss rates for each sector in each country. - Assesses how banks are directly affected by borrower support measures (debt moratoria, credit guarantees, deferred bankruptcy proceedings), complementing public information with IMF desk surveys. - Analytical approach contrasts with other exercises by using the January 2021 WEO projections and by modeling dynamic bank balance-sheet evolution (two-period model with new lending conditioned on prior-period capital). ### Key findings on capital and policy effects - Aggregate CET1 capital ratio of euro area banks directly supervised by the ECB increased by almost 30 basis points from end-2019 to end-September 2020. - Although the impact of the crisis on banks has been limited so far, capital pressures are expected to rise during a protracted recovery. - Continued borrower-support and capital conservation measures have allowed banks to slowly absorb rising impairments without a significant change in capital ratios to date. - Vulnerabilities may emerge as policy support measures expire, potentially causing a rise in impairments and a surge of bankruptcies due to declining debt service capacity of nonfinancial corporations and households. - A more sluggish recovery than in the baseline, or a premature phaseout of support measures creating “cliff effects,” could amplify deleveraging pressures on weaker banks and banks most exposed to vulnerable sectors. ### Pre-pandemic bank health (selected indicators) - Common Equity Tier 1 (CET1) capital: nearly 15 percent of risk-weighted assets in 2019; nearly doubled from about 7 percent in 2007; well above prudential minimum of 4.5 percent (plus bank-specific Pillar 2 requirement). - Non-risk weighted leverage ratio (Tier 1 capital as percent of total assets): around 6 percent on average in 2019; about twice the minimum threshold of 3 percent that will be binding from June 2021. - Nonperforming loans (NPLs): peaked at 7½ percent of gross loans after the GFC in 2013; declined to about 3 percent of gross loans by 2019. - Profitability: return on assets (ROA) of many European banks declined since the GFC; euro-area banks’ ROA was below peers in other advanced economies in 2019. - Market valuation: a quarter of euro-area banking system assets were trading with price-to-book ratios below 0.4 in 2019 (share increased further in 2020). ### Cross-country heterogeneity - Wide variation in CET1 capital ratios and profitability (ROA) across countries. - Some vulnerable euro-area economies (for example, Cyprus and Greece) continue to carry a large stock of NPLs (high share of Stage 3 loans) and a high share of Stage 2 loans, indicating loans on the cusp of becoming NPLs. - The disproportionately high share of Stage 2 loans covered by debt moratoria suggests latent asset quality pressures. ### Banks’ vulnerability to the pandemic (exposures and firm stress) - More than 60 percent of banks’ corporate exposures are to sectors highly affected by the pandemic (especially accommodation and food services, real estate and retail trade, and to a lesser extent construction and transportation). - Bank lending to firms in these highly affected sectors is about 200 percent of Tier 1 capital on average and exceeds 250 percent in some countries (examples cited: Finland, Germany, Greece, and Italy). - More than half of bank lending is to households, especially mortgages, which are increasingly affected by adverse income and employment prospects. - IMF Regional Economic Outlook (REO) October 2020 estimates (top-down): absent supportive policies, 30 to 40 percent of firms in advanced Europe and up to 50 percent in emerging Europe would face liquidity gaps in 2020; share of insolvent firms (negative net worth) could rise to 20 to 30 percent in the median advanced and emerging European economy. - Policy support from European authorities is expected to reduce the number of illiquid firms in Europe by about two-thirds and, by design, reduce the number of insolvent firms by a lesser amount. ### Paper structure (contents by chapter) - Chapter 2: financial soundness of European banks before COVID-19. - Chapter 3: stylized facts on European banks’ vulnerability to the pandemic-related shock. - Chapter 4: summary of key policy measures that have direct or indirect impact on banks. - Chapter 5: description of the analytical framework. - Chapter 6: results. - Chapter 7: extension to a wider set of European banks. - Chapter 8: conclusions with policy implications. *Source: Introduction (chapter), chwebfea - Introduction, IMF staff paper.* ### 2. Euro Area: Large Banks’ Exposure to HH and NFCs ### 2. Euro Area: Large Banks’ Exposure to HH and NFCs ### Regulators’ capital and prudential relief - Regulators provided substantial capital and liquidity relief to strengthen banks’ capacity to absorb losses while continuing to lend: - Banks were allowed to use their combined capital buffers and were temporarily allowed to operate below the level of capital required under Pillar 2 guidance and the liquidity coverage ratio. - Prudential authorities temporarily granted flexibility in the classification and greater clarity on the provisioning of loans backed by public support measures (EBA 2020e). - Temporary measures were enhanced by lower countercyclical capital buffers set by national macroprudential authorities and complemented by capital conservation measures such as restrictions on dividend distribution and share buy backs, supporting bank lending. - Targeted supervisory guidance and timeline notes in the source: - On March 20, the ECB introduced supervisory flexibility regarding the classification of debtors as “unlikely to pay” on public guarantees granted. - On March 25, EBA published guidance on the definition of default, forbearance and the application of IFRS 9 in the context of COVID-19. - EBA published Guidelines on Legislative and Non-legislative Moratoria on Loan Repayments on April 2, 2020 (EBA/GL/2020/02). ### Policy measures for firms and households - Labor support: - Several economies expanded short-time work schemes that use public funds to supplement up to 70–80 percent of employees’ pay for hours not worked. - Direct support instruments: - Cash grants, tax deferrals, and expanded direct lending to firms were provided in several countries. - Loan guarantees: - According to a survey of IMF “desk economists” for 44 European countries, 39 countries offer pandemic-related guarantee programs. - Seven countries offer an overall guarantee envelope well in excess of 10 percent of GDP. - Public guarantee programs are often targeted at SMEs; examples given: Norway, Portugal, and Serbia programs exclusively for SMEs; in Slovenia, Switzerland, and the United Kingdom more than 90 percent of the total guarantee envelope is for SMEs. - Guarantees tend to have a coverage ratio of 70 to 90 percent. - For EU member countries, 100 percent government guarantees are allowed only for loans up to €800,000 according to State Aid rules; on January 28, 2021, the European Commission announced an increase in the ceiling to €1.8 million per company. - Impact on bank lending: - Euro area bank lending growth increased from 5 percent (year over year) at the beginning of 2020 to nearly 9 percent (yoy) in May amid strong precautionary cash demand. - New lending growth to firms fell to about 6½ percent (year over year) by October, a rate double the recent multi-year average. - Government guarantees lower capital costs for banks because risk-weights on such loans are lower. ### Debt service moratoria and insolvency moratoria - Debt service moratoria: - Our survey indicates that 38 European countries introduced debt service moratoria. - Moratoria can be legislative or non-legislative (private-sector initiatives). - In most countries, moratoria were provided for both households (mortgage and consumer loans) and businesses. - The duration of moratoria originally ranged from three to six months and was extended in some countries into 2021. - Debt service moratoria reduce immediate cash receipts and therefore can entail a loss of interest income for banks (accounting on an accrual basis still counts deferred but not cancelled interest as interest income; however, moratoria may delay cash receipt beyond the stress test horizon and thus impose an economic loss). - Prudential authorities emphasized flexibility in classification of loans eligible for debt repayment relief; EBA temporary guidelines advised supervisors and banks to “look through” the transitory systemic shock, suspend days-past-due automaticity in classifying loans as forborne or defaulted, and focus on identifying borrowers with longer-term financial difficulties. - EBA temporarily lapsed these guidelines in mid-September but reinstated them in December while emphasizing timely recognition of credit risk. - Insolvency moratoria: - Most countries adopted insolvency moratoria for about half a year from the end of March 2020 onward. - Deferred bankruptcy proceedings likely delayed loan write-offs due to corporate default; potential cliff-edge risks may arise once insolvency moratoria fully expire, leading to suddenly rising bankruptcies and adverse impacts on bank capital. - Figure 12 summary in the source: Median duration of COVID-19 pandemic-related insolvency moratoria — EA median: 5.7 (months); Other EU/EEA median: 0 (months). ### Top-down stress testing approach and data sources - Approach: - A standard top-down stress testing approach using publicly available data was adopted to assess banks’ capital buffers against the pandemic shock. - The exercise focuses on declines in bank capital ratios primarily due to credit risk; it does not include interconnectedness within the financial sector or detailed liquidity stress tests. - Data sources: - EBA’s 2020 Transparency Exercise for 117 large European banks (main exercise), with results focused on 90 euro area banks (the “EBA sample”). - Bank-specific time-series data for 2008–19 from FitchConnect on assets, capital levels, profitability (ROA), NPLs, and lending growth to estimate empirical links between macroeconomic performance and ROA, ROA components, loan growth and NPLs. - Macroeconomic baseline projections (GDP growth, unemployment) from the IMF World Economic Outlook database. - S&P Global Market Intelligence data to expand coverage to smaller euro area banks and banks in non-euro area European economies; sample coverage summarized in Figure 13. ### Capital impact channels and modeling specifics - Three channels through which the pandemic affects bank capital: 1. Profitability channel — lower operating income and higher provisions reduce retained earnings and CET1 accumulation. 2. Asset channel — increased write-offs from corporate bankruptcies and new lending alter asset levels. 3. Risk-weight channel — higher credit risk weights increase risk-weighted assets. - Interaction with policies: - Demand-side supports (credit guarantees, moratoria, grants) and supply-side supports (prudential measures, deferred losses) influence the propagation of shocks through the three channels. - Profitability channel quantitative sensitivities (satellite model results): - Every 1 percentage point decline in real GDP growth reduces the ROA of large euro area (SSM) banks by 27 basis points. - Every 1 percentage point increase in the unemployment rate reduces ROA by 21 basis points. - A 1 percentage point increase in the last period’s NPL ratio reduces ROA by 15 basis points. - Asset channel specifics: - Write-offs: the pandemic is expected to lead to a surge in bankruptcies and loan losses above historical averages; impaired corporate loan amounts are derived from sector- and country-level mappings in the source analysis (Chapter 3 of the October 2020 Regional Economic Outlook: Europe). - Lending pace: corporate credit demand was strong during the pandemic to fund working capital needs; supply constraints are assumed to bind loan growth in the current crisis. The empirical model links changes in lending to initial bank capital ratios. ### Key numeric markers from figures and sample descriptions - Top visual annotations: - Average = 234 - Average = 196 - Most affected banks (34% of total assets) - Coverage and samples: - EBA coverage data: 117 EU/EEA banks, of which 90 are the largest euro area banks (~SSM). - Extended coverage: 467 European banks (covering all European IMF member countries, except Andorra, Israel, Kosovo, Moldova, and North Macedonia), of which 138 are euro area banks; includes the “EBA coverage.” - Lending growth snapshots: - Lending growth: 5 percent (yoy) at beginning of 2020 → nearly 9 percent (yoy) in May 2020. - New lending growth to firms: about 6½ percent (yoy) by October 2020. - Survey counts and envelopes: - 39 of 44 European countries surveyed offer pandemic-related guarantee programs. - 38 European countries introduced debt service moratoria. - Seven countries offer a total guarantee envelope well in excess of 10 percent of GDP. - Guarantee coverage ratios and ceilings: - Typical guarantee coverage ratio: 70 to 90 percent. - EU State Aid rule ceiling for 100 percent guarantees: €800,000 per company (European Commission announcement on January 28, 2021 increased ceiling to €1.8 million per company). *Italic: Source — chwebfea - 2. Euro Area: Large Banks’ Exposure to HH and NFCs (excerpt) from the provided PDF content.* ### 0.6 percentage point decline in credit growth (Annex Table 2.4). In turn, ### chwebfea - 0.6 percentage point decline in credit growth (Annex Table 2.4). In turn, ### Channels affecting the capital ratio - Profitability channel: higher provisions and impairment charges reduce net income and CET1. - Asset channel: lower credit growth directly reduces asset growth in the next period, mechanically generating an increase in the capital ratio; but higher write-offs reduce capital. - Risk-weight channel: changes to credit risk weights (via updated model-implied probability of default (PD)) affect risk-weighted assets and thus the capital ratio. - Empirical balance: a negative shock that increases write-offs and reduces new lending has an ambiguous net effect on capital ratios (falling due to higher write-offs, rising due to lower lending). ### Model treatment and parameter extraction - One-year PD is extracted from credit risk weights (EBA Transparency Exercise 2020) using the internal ratings-based approach (IRB) formula in the Basel framework. - The increase in loan loss provisions is estimated from sub-components of the ROA satellite model. - Uniform but asset class-specific recovery rates and maturities are applied for each country based on EBA’s quarterly risk parameters (EBA 2021b) and its report on insolvency frameworks (EBA 2020c). - The PIT/TTC distinction: the approach generates a point-in-time (PIT) measure of default risk (linked to changes in provisions under IFRS-9). Regulatory capital calculations require through-the-cycle (TTC) parameters, which would result in lower PD estimates under stress. - Higher provisions reflect the increase of general default risk of all exposures, including additional impairments of corporate loans due to corporate insolvencies. ### Key estimated sensitivities (ROA impacts) — Table 2 - Real GDP growth → change in ROA (in pp, by sample): 0.270***; 0.141***; 0.197**; 0.0752** - Unemployment rate → change in ROA (in pp, by sample): –0.208**; –0.0789*; –0.171***; –0.0557 - NPL ratio (lagged) → change in ROA (in pp, by sample): –0.0155; –0.0391***; –0.0418**; –0.0295* - Note: EA = euro area; EA, ex. SSM = euro area non-SSM banks; EBA = European Banking Authority; NPL = nonperforming loan; ROA = return on assets; SSM = sample of large euro area banks supervised by the Single Supervisory Mechanism. ***p < 0.01, **p < 0.05, *p < 0.1. ### Policy measures, transmission, and model treatment - Demand-side borrower support (grants, tax deferrals/exemptions, wage subsidies) reduce corporate default probability and write-offs (“asset” channel). - Credit guarantees reduce expected losses; debt moratoria reduce borrower default risk, slowing the rise in provisions (“profitability channel”) and risk weights (“risk channel”). - Debt moratoria also weigh on bank profitability via deferred or lost interest income and caps on interest rates of some guaranteed loans. - Supply-side prudential measures (capital buffer relief such as countercyclical capital buffer and systemic risk buffer, and delayed provisions if no automatic reclassification under moratoria) delay loss recognition and dampen provisions and risk-weight increases. - Assumed sluggish adjustment of provisioning for loans subject to moratoria relative to expectations (based on satellite model estimations; Annex 2). - Examples of pre-pandemic buffers: counter-cyclical capital buffer in Ireland was 1 percent, in France and Germany was 0.25 percent; Denmark’s counter-cyclical buffer reduced to zero at pandemic onset then raised twice to 1.5 percent in June 2020 and 2.0 percent in December 2020. ### Corporate default shock construction and mapping - Corporate exposure data from EBA Transparency Exercise; sectoral loss rates based on Chapter 3 of October 2020 REO: Europe. - Two-step mapping: 1. Determine within-sector, debt-weighted share of firms likely to become illiquid and insolvent (sector-specific default rates), with and without policy measures. The share of firm debt at risk of default ranges from 2 percent in Germany to 7 percent in Italy. 2. Map these default rates to each bank’s corporate exposure by sector to determine additional corporate default losses, then translate to bank-level write-offs after accounting for existing loan loss coverage and average recovery values (LGD proxies). - Policy impact on corporate losses: policies reduce loan losses by half in Germany and by two-thirds in France (example results). - Timing: on average, only 15 percent of the calculated corporate shock occurs in 2020, with the remainder materializing in 2021 (reflecting deferred insolvency proceedings). ### Stress test scenarios and macro assumptions - Two macro scenarios calibrated to January 2021 World Economic Outlook (WEO) Update: - Baseline: sharp recession in 2020 followed by partial recovery in 2021 (Figure 17). - Illustrative adverse: slower recovery in 2021 if vaccinations proceed slower and containment measures are more stringent/longer-lasting. - Note: ECB March 2021 projections suggest euro area GDP fell by 6.9 percent in 2020. - Example numeric indicators shown in Figure 17: –7.3; cumulative values and short-term growth figures as specified in the scenario figures. ### Stress test results (baseline) — solvency outcomes and channel decomposition - Aggregate CET1 capital ratio for large euro area banks declines from 14.7 to 13.0 percent by the end of 2021 (baseline, accounting for granular borrower support measures). - Capital erosion decreases by about one-third if banks continue to refrain from dividend distributions in 2021. - Without supportive policies, capital erosion would be 1.7 percentage points larger (roughly double the scale of capital depletion). - Channel contributions with supportive policies: - Profitability channel reduces CET1 by 0.7 percentage points. - Asset channel impact: 0.5 percentage points (reported as slightly smaller than profitability channel). - Risk-weight channel impact: 0.5 percentage points. - Without policies: lower profitability explains most capital depletion and higher contribution from increased credit risk weights; asset channel may act to slightly increase the capital ratio (reduced lending mechanically lowers denominator, dominating write-off impact). ### Additional quantitative and calibration notes - The analysis covers three channels: profitability (net interest income and provisions), nominal assets (net lending and charge-offs after reserves), and risk exposure (changes in credit risk weights). - Policy types modeled include: debt repayment relief (moratoria), public credit guarantees, delayed insolvency proceedings, dividend restrictions (only in 2020). - Regulatory reference points shown in figures: regulatory minimum (4.5%) and MDA (9.1%). *Source: IMF staff.* ### 1. Aggregate CET1 Capital Ratio (before and after shock) ### 1. Aggregate CET1 Capital Ratio (before and after shock) ### Key findings - Aggregate CET1 ratio (end-2019): 14.7 - Projected CET1 ratio (end-2021, baseline with policy measures): 13.1 - Projected CET1 ratio (end-2021, baseline without policy measures): 11.5 - IMF exercise: aggregate capital impact over two years of 1.7 percentage points (with policy measures) - Adverse scenario: additional CET1 capital ratio decline of 1 percentage point by end-2021 even with current policy measures - Regulatory minimum for CET1: 4.5 percent - Capital Conservation Buffer (CCB): 2.5 percent (implied hurdle with regulatory minimum = 7 percent) - Average MDA (maximum distributable amount) threshold: about 9 percent of risk-weighted assets - About 14 percent of the largest (90) euro area banks are likely to breach their MDA (baseline, without policy support) - With policies, the aggregate CET1 ratio is lifted to above the 7 percent threshold in all countries (baseline) ### Country- and bank-level results - Several banks would breach the 7 percent hurdle rate without policy support. - In the illustrative adverse scenario: - More than five percent of the large euro area banks (6 banks) would see CET1 drop below the MDA threshold (with current policy measures). - Without policies, more than a quarter of larger euro area banks (25 banks) would breach the MDA threshold. - Potential capital need estimates: - If ~14 percent of large banks breach MDA without policies: required capital ~€25 billion (or 1.7 percent of reported CET1 capital at the end of 2019). - Without policies under the adverse scenario: capital need nearly €47 billion (or 3.1 percent of reported CET1 capital at the end of 2019). - Banks concentrated at risk under adverse scenario: Italy, Portugal, and Spain (sample coverage and figures referenced). ### Breakdown and drivers of CET1 changes - Components affecting CET1 under stress: change in assets, profitability impact, change in risk weights. - With policies: - Average increase in risk weights among euro area banks: about 1.7 percent. - Without policies: - Risk weights increase by more than 4 percent and exhibit slightly larger upside dispersion. - Policy measures included in the analysis: debt repayment relief (moratoria) for businesses and households, public/corporate credit guarantees, deferred insolvency proceedings, and dividend restrictions (only in 2020). - Profitability impact and other channels are explicitly modeled (net interest income, provisions, asset write-offs, and changes in credit risk weights). ### Role and effects of policy measures - Policy support materially improves bank solvency and reduces dispersion in post-shock CET1 ratios across banks. - With policies: - Smaller hit to capital ratios and reduced dispersion, especially in the lower half of the CET1 distribution. - Dampening effect operates mainly through limiting increases in risk weights. - Without policies: - Higher dispersion in CET1 outcomes and larger increases in risk weights. - Policy implications: - Careful communication of buffer usability and capital relief policies is crucial to dampen capital hits, especially under adverse scenarios. - Dividend restrictions and flexibility in use of hybrid capital are important near-term policy tools noted in the analysis. ### Adverse scenarios and sensitivity - Adverse scenario assumption: cumulative GDP growth over 2020–21 is 1.2 percentage points below baseline (illustrative adverse). - Under a more severe adverse scenario (e.g., ECB-style scenario), additional output loss would be larger and capital depletion materially greater. - Sensitivity: - Increasing severity of the adverse scenario would raise the number of banks at or below the MDA threshold; a more severe adverse case could result in about a quarter of banks being at or below MDA even with policy measures. - Figure-based sensitivity analysis shows CET1 ratios respond nonlinearly as the multiplier of cumulative GDP shock increases (scenario severity). ### Capital replenishment and timelines - End-2021 CET1 under stress-test baseline: 13.1 percent (gap relative to pre-crisis 14.7 percent = 1.6 percentage points). - Time to restore precrisis CET1 = 14.7 percent through profits after end-2021 under different return-on-assets (RoA) assumptions: - Precrisis RoA = 0.38 percent → 2.6 years to restore precrisis capitalization - Half precrisis RoA = 0.19 percent → 5.2 years to restore precrisis capitalization - Required RoA = 1.06 percent → 1 year to restore precrisis capitalization (illustrative required profitability) - Implication: absent a material rise in profitability, organic replenishment of capital buffers will take time; if earnings capacity is subdued (about half historical average) replenishment could take more than five years. - Hybrid capital: - Close to one-quarter of the capital base of the large euro area banks comprised various forms of hybrid capital at end-2020; hybrid capital could be used to help replenish aggregate capital levels. ### Comparison with ECB Vulnerability Analysis - ECB Vulnerability Analysis (July 2020) of 86 largest euro area banks: - Baseline (“central scenario”) projected decline of economic activity by 8.7 percent in 2020, then growth of 5.2 percent in 2021 and 3.3 percent in 2022. - Cumulative output loss in ECB central scenario during 2020–21: 6.0 percent (similar to IMF baseline 5.8 percent decline). - ECB found CET1 declines of 1.9 percentage points in baseline over a three-year horizon. - Comparison with IMF results: - IMF exercise finds a 1.7 percentage point capital impact over a two-year period (with policy measures). - IMF projects a smaller increase in risk-weighting of credit-sensitive assets than the ECB (conditional on effective policy measures). - Differences partly reflect IMF’s more comprehensive coverage of supervisory and fiscal relief measures (debt repayment relief and deferred bankruptcy proceedings included in IMF analysis but excluded from ECB’s Vulnerability Analysis). *Source: chwebfea - 1. Aggregate CET1 Capital Ratio (before and after shock).* ### Box 2. Comparison of Stress Test Results with the ECB Vulnerability Analysis ### Box 2. Comparison of Stress Test Results with the ECB Vulnerability Analysis ### Key results and scenario comparisons - Coverage expanded from 90 euro area banks (“EBA coverage”) to 468 banks in 40 countries. - Main transmission channels considered: profitability (net interest income and provisions), nominal assets (net lending and write-offs after reserves), and risk exposure (changes in credit risk weights). - IMF two-year cumulative GDP deviations and ECB comparisons (Box Figure 2.1 / Box Figure 2.2) show large negative growth shocks in 2020–2021 under pandemic scenarios (figures in source: –7.3; 4.2; –3.1; 3.1; –4.2; –8.7; 5.2; –3.5; –5.8; –6.0; –5.1; –4.3; –5.0 as presented in the Box Figures). ### Findings from the expanded sample (468 banks) - Sample composition and assumptions: - Original EBA sample: 90 euro area banks. - Expanded sample: 468 banks in 40 countries, including smaller euro area banks and banks in non-euro area advanced economies and CESEE countries. - For banks not in the EBA Transparency Exercise, sectoral exposure is proxied by the asset-weighted average sectoral exposure of EBA-covered banks in the same country or neighboring country. - Sensitivities (profitability, NPLs, loan growth) estimated with subsample-specific coefficients from the satellite model. - Capitalization and solvency outcomes: - End-2019 CET1 capital ratios: expanded-sample euro area banks 14.9 percent; larger euro area banks in EBA Transparency Exercise 14.7 percent. - Projected CET1 capital ratios under stress are roughly the same for expanded sample and EBA-covered large banks. - Under baseline scenario, about 9 percent of all euro area banks in the expanded sample are likely to breach their MDA threshold if policies do not operate as expected; required new capital to avert market pressure: €26 billion (or 1.7 percent of reported CET1 capital at the end of 2019). - With supportive policies, about 3 percent of all euro area banks in the expanded sample would struggle to clear the MDA hurdle rate, generating a capital shortfall of €0.6 billion. - In the adverse scenario, the number of banks likely to fall below the MDA threshold doubles relative to baseline (with and without policies). - Non-EU CESEE banks: - Lower CET1 capital ratio at end-2019 and higher projected capital erosion than euro area banks. - Projected CET1 capital ratios: 10.8 percent (baseline) and 10.5 percent (adverse). - Mitigating policies lift CET1 by 0.4 percentage point in non-EU CESEE banks, versus around 1.3 percentage points for euro area banks covered by the EBA Transparency Exercise. - For non-EU CESEE banks, the risk weight channel explains most of the decline in CET1 capital ratios (reflecting low asset quality and prior NPL buildup). - Channel contributions and policy sensitivity: - For euro area and EU banks in the expanded sample under the baseline policy scenario, contributions from profitability, assets, and risk weights are roughly balanced. - Absent policies, capital erosion from the profitability channel becomes about three times larger, and twice as large from the risk weight channel in the baseline macro scenario. - Profitability channel remains relatively more policy-responsive; absolute magnitudes increase under adverse scenarios. ### Macroeconomic and credit supply implications - Most banks entered the pandemic with sizeable capital buffers; deterioration in asset quality will hurt profitability, especially for banks with large exposures to vulnerable sectors. - Policy support during initial lockdowns limited unemployment and bankruptcies, allowing banks to absorb impairments without major CET1 declines. - Risks if the crisis persists: - Subdued activity and delayed reopening would worsen borrower liquidity, increase debt overhang, and produce higher loan loss provisions and credit losses. - Diminishing capacity to lend could weigh on consumption and investment when financing is most needed. - Estimated effects on lending: - Under the baseline scenario, capital constraints could reduce lending growth by about 1.6 percentage points next year. - If policy measures do not fully operate as expected, credit growth could slow by about 3 percentage points—corresponding to the average credit growth of large euro area banks in 2019. - Under the adverse scenario, capital erosion is much larger after support measures expire, leading more banks to de-leverage as CET1 ratios approach the MDA threshold. ### Policy implications and recommended strategy - Adopt a multi-pronged strategy focusing on borrower support, banks’ balance sheet resilience, and a careful rollback of temporary measures. Key borrower-support considerations: - Debt repayment relief: - Tailored bank-led debt-service relief should remain central to help stressed but potentially viable borrowers. - Broad, general moratoria were effective initially but are unsustainable long term because they defer accrued interest income, pressure net operating income, and can distort loan classification and provisioning. - As recovery proceeds, moratoria should be better targeted to illiquid but viable firms and vulnerable households; moratoria should be extended only if they do not risk distorting classification and provisioning, and banks should be encouraged to restructure debts of illiquid but likely viable borrowers. - Credit guarantees: - Public sector credit guarantees can be preferable to mandatory blanket moratoria as a targeted tool. - Guarantee design should align banks’ incentives with public interest—ensuring “skin in the game” so banks select and service borrowers appropriately and limit losses beyond addressing market failures. *Italic: IMF staff calculations and source figures as presented in Box 2 of the referenced chapter.* ### 2. Share of Nonperforming Loans in Amount of Loans ### 2. Share of Nonperforming Loans in Amount of Loans under Moratoria ### Nonperforming loans under moratoria — snapshot - Country-level values shown include: 6.4; Regional median: 7.7; 2.3; 5.7. - Country ISO labels appear in figures (examples in source): PRT, CYP, IRL, ITA, BEL, GRC, ESP, AUT, FRA, NLD, DEU, HUN, ROU, SVN, MLT, SVK, LUX, LVA, EST, FIN, LTU, HRV, GBR, CZE, POL, ISL, SWE, CHE, BGR, DNK, NOR. - Debt repayment relief measures referenced include moratoria for businesses and households, corporate credit guarantees, delayed insolvency proceedings, and dividend restrictions (only in 2020). ### Capital relief and conservation measures — findings and recommendations - Finding: Many banks have been reluctant to dip into capital buffers since effective hurdle rates, such as the MDA, are much higher than the current prudential minimum (ECB 2020b). - Finding: Supervisors encouraged banks to use their capital buffers (Figure 8; ECB 2021c), but any breach of the combined buffer requirement (a significant part of the MDA) will lead to restrictions on dividend distributions and coupon payments on hybrid capital. - Recommendation: Supervisors need to clearly convey to banks and investors the extent to which capital buffers can be used to avert market pressures. - Recommendation: Adopt a realistic timetable for replenishing capital buffers; maintain restrictions on dividend payouts and share buybacks until the recovery is well underway to allow gradual rebuilding of capital and liquidity buffers without impairing lending capacity. - Timing and guidance noted: current supervisory guidance states that capital buffers can be used through the end of 2021, and that the capital add-on under Pillar 2 Guidance does not need to be replenished until after 2022 (ECB 2021b). - Caution: Banks may be cautious about using buffers because they can take a long time to replenish (Figure 28). An overly rapid replenishment strategy could discourage buffer use and slow the recovery (Borsuk, Budnik, and Volk 2020). ### NPL management and bank resolution frameworks — findings and recommendations - NPL management - Recommendation: Normalize and clearly communicate prudential standards as recovery takes hold to incentivize timely recognition of problem assets via greater balance sheet transparency and upgraded reporting. - Recommendation: Enhance NPL monitoring to ensure banks can adequately provision for impaired loans. - Recommendation: Foster development of secondary markets for distressed assets to facilitate disposal of NPLs, particularly for smaller banks. - Consideration: Asset management companies (AMCs) could help offload NPLs in some countries and loan types, but must proceed with care and safeguards because NPL sales during stress likely entail losses borne by banks’ shareholders or governments. - Insolvency proceedings - Finding: Deferral of insolvency proceedings has delayed defaults but created legacy risk of pent-up creditor claims and reduced asset recovery prospects. - Historical benchmark: Bankruptcy proceedings have taken about one to three years historically, with substantial heterogeneity among countries. - Recommendation: Provide EU-level benchmarks for upgrading insolvency regimes in Member States to reduce time and cost of insolvency proceedings; fast-track court procedures and address administrative constraints. - Recommendation: Put in place efficient out-of-court workouts with separate tracks for firms, SMEs, and households. - Bank capital planning and resolution - Finding: If banks experience significant depletion of capital buffers and market conditions are unfavorable for raising fresh capital, taxpayer-funded capital injections might be necessary. - Recommendation: Use the current system-wide stress test (EBA 2020e, 2021a), expected to be completed in July 2021, to assess potential recapitalization needs under a realistic adverse scenario. - Recommendation: Precautionary recapitalizations could use flexibility provided by the Temporary State Aid Framework (EC, 2020). - Recommendation: Use stress-test results to challenge banks’ capital projections in SREP, foster consistency in risk assessments, and promote prudent provisioning policies. ### Tackling chronic low profitability — findings and recommendations - Finding: Over the medium term, structurally low profitability will limit banks’ ability to restore capital buffers organically; earnings capacity is likely to remain subdued. - Finding: An increasing number of banks are reporting earnings below their cost of capital (Figure 32; IMF 2021b). - Finding: Investments in digital technologies to reduce structural margin and cost pressures add to short-term expenses. - Recommendation: Further consolidation of banking groups through domestic and cross-border M&A could improve efficiency and cross-border risk sharing. - Recommendation: Eliminate remaining prudential and legal obstacles to cross-border integration of banking activities. - Note: Recent ECB guidance on use of supervisory tools to facilitate sustainable consolidation is identified as timely (ECB 2021b). ### Key regulatory numeric thresholds and stress-test references - Regulatory minimum CET1 capital ratio: 4.5% - Capital conservation buffer (CCB): 2.5% - MDA (weighted average) examples in figures: MDA (10.6%); MDA (9.1%) - Supervisory guidance on buffer use: can be used through the end of 2021; Pillar 2 Guidance add-on replenishment not required until after 2022. - Stress-test milestone: system-wide stress test expected to be completed in July 2021. - ECB recommendation on dividend payouts: generally encouraged to refrain from or limit shareholder payouts until September 2021. - Historical duration of bankruptcy proceedings: about one to three years (substantial heterogeneity across countries). *Source: chwebfea - 2. Share of Nonperforming Loans in Amount of Loans under Moratoria (from COVID-19: HOW WILL EUROPEAN BANKS FARE? chapter).* ### Annex Figure 1.7. European Banks: Changes in Capital Ratio and GDP Growth, Extended Coverage ### Annex Figure 1.7. European Banks: Changes in Capital Ratio and GDP Growth, Extended Coverage ### Key concepts and methodology - CET1 = common equity Tier 1; CCB = capital conservation buffer; MDA = maximum distributable amount (weighted average). - The analysis covers three channels affecting the capital adequacy ratio under stress: profitability (net interest income and provisions), nominal assets (net lending and write-offs after reserves), and risk exposure (changes in credit risk weights). - The boxplots' grey shaded area shows the interquartile range (25th to 75th percentile), with whiskers at the 5th and 95th percentile of the distribution. - Policy measures modeled in scenarios include: debt repayment relief (moratoria) for businesses and households, corporate credit guarantees, delayed insolvency proceedings, and dividend restrictions (only in 2020). ### Baseline vs Adverse scenarios — cross‑sample comparison - Baseline and Adverse scenarios are reported across sample groups: Euro area, Non‑EU advanced, EU, Non‑EA CESEE, Non‑EU CESEE, Total sample. - The figure maps change in CET1 capital ratio against cumulative GDP growth (2020–21) for each sample grouping. ### Aggregate CET1 capital ratios and projections (selected panels across figures) - Euro Area (Extended Coverage, baseline / adverse panels shown across figures): - CET1 ratio (end-2019): 14.9 - Baseline projected CET1 ratio (end-2021, with policy measures shown in related panels): 11.8; alternate panel values shown include 13.1 and other subgroup averages. - Adverse scenario panels show aggregate CET1 ratios with projected declines to values such as 12.1 and 10.9 in various subgroup breakdowns. - EU Banks (Extended Coverage): - CET1 ratio (end-2019): 15.3 - Baseline projected CET1 ratio (end-2021, with policies): 13.5 and subgroup projected values 12.0. - Adverse scenario projected CET1 ratios (end-2021): 12.5 and 11.1 in reported panels. - Non‑EU CESEE Banks: - CET1 ratio (end-2019): 12.8 - Baseline projected CET1 ratio (end-2021, with policies): 10.8 and 10.4 in different panels. - Adverse scenario projected CET1 ratios (end-2021): 10.5 and 10.1 in reported panels. - All Banks (Extended Coverage, aggregate): - CET1 ratio (end-2019): 15.3 - Baseline projected CET1 ratio (end-2021, with policies): 13.1 and subgroup projected value 11.6. - Without policy measures projected CET1 ratio (end-2021) examples: 12.8 and 11.6 in panels. ### Breakdown of drivers of CET1 changes (examples reported in panels) - Change in assets (nominal asset channel) — reported as separate component in panel breakdowns. - Profitability impact (net operating income after provisions and losses) — reported as separate component; policy measures reduce provisioning for guaranteed loans and allow loss forbearance under moratoria, affecting this component. - Change in risk weights (credit risk exposure) — reported as separate component; increases reflect higher unexpected losses and additional provisions. - Example numeric component values shown in panels (select reported values): - Euro Area baseline panels show component contributions including 0.6, 0.8, 0.4, 2.6, 1.0, –0.5 (as labelled in a breakdown). - EU baseline panels show component contributions including –0.5, 1.1, 2.7, 0.6, 0.4, 0.8. - All Banks baseline panels show component contributions including 0.7, 1.0, 0.6, –0.6, 2.9, 1.4. - Non‑EU CESEE baseline panels show component contributions including –0.04, 0.3, 0.1, 0.2, 1.7, 2.1. - Adverse scenario panels report additional component magnitudes, for example values such as –0.7, 3.4, 1.3, 0.9, 1.3, 0.7 in Euro Area adverse breakdowns. ### Policy effects and conditional outcomes - With policy measures (moratoria, guarantees, delayed insolvency proceedings, dividend restrictions in 2020) projected CET1 ratios are consistently higher than without policies across sample groups in both Baseline and Adverse scenarios. - Examples of policy impacts in panels: - Euro Area panels compare current (end-2019) CET1 of 14.9 to projected (end-2021) CET1 values under “With policy measures” and “Without policy measures” labels; reported projected values include 14.7, 13.1, 11.8, 11.5 depending on subgroup. - EU and All Banks panels display similar comparisons, showing smaller declines when policies are applied. ### Stress channels and modelling notes - Corporate write-offs and net lending are modeled as arising from illiquid and insolvent firms, weighted by outstanding debt and mapped to sector-by-sector corporate exposures of sample banks. - Net profitability impacts of policies incorporate: lower provisions for guaranteed loans to solvent corporates, loss forbearance on eligible loans under moratoria, decline in interest income due to duration of debt moratoria (households and businesses), and changes in net operating income after general provisions and losses on other noninterest income due to lower GDP growth and higher unemployment, including impairment charges for noncorporate exposures. - Increase of credit risk weights results from higher unexpected losses (derived from the increase of default risk implied by the projected increase of general provisions) and additional specific provisions for additional corporate loan losses. *Sources: EBA; ECB; ESRB; FitchConnect; S&P Market Intelligence; and IMF staff estimates.* ### Annex Figure 1.15. All Banks (Extended Coverage): Solvency Stress Test—Adverse Scenario ### Annex Figure 1.15. All Banks (Extended Coverage): Solvency Stress Test—Adverse Scenario (Percent) ### Overview of the solvency stress test - Purpose: Project changes in banks’ Common Equity Tier 1 (CET1) capital ratio from end-2019 through end-2021 under an adverse COVID-19 scenario, with and without policy measures. - Three channels considered: - Profitability channel (retained earnings via ROA and its components). - Asset channel (write-offs/impairments and net change in assets). - Risk channel (higher credit risk-weights reflecting rising default risk). - Starting point: CET1 ratio at end-2019. ### Aggregate CET1 outcomes (figural values preserved) - Charted CET1 ratio levels (before and after shock) include numeric markers present in figure layout: 18, 16, 14, 12, 10, 8, 6, 4, 2, 0 (vertical scale repeated). - Specific CET1 ratio values and projected changes shown in breakdowns (with policies and without policies): - CET1 ratio (end-2019) reported values in breakdown: 15.1; 15.3; 11.6; 12.0; 9.9; 10.6; 15.3; 12.0. - Change components and projected (end-2021, proj.) values: 1.0; 1.5; 0.8; 3.7; 1.7; –0.7. - Labels in figure: "Current (end-2019)"; "With policy measures"; "Projected (end-2021)"; "Without policy measures". ### How policy measures are incorporated - Bank-facing measures considered: - Financial sector policies including greater supervisory and regulatory flexibility and capital easing to support lending. - Credit-demand measures such as debt moratoria and credit guarantees. - Fiscal and other measures considered for indirect effects on bank asset quality: - Grants, wage subsidies, commercial rate reductions, tax deferrals. - Mechanisms of policy impact on the three channels: - Profitability channel: lower provisions for guaranteed loans to solvent firms; loss forbearance on eligible moratoria loans; decline in accrued interest income during moratoria. - Asset channel: additional write-offs from projected insolvency of illiquid/insolvent firms (single-factor corporate shock), plus impairment charges for noncorporate exposures. - Risk channel: public sector guarantees reduce marginal credit risk weight of new corporate loans to solvent borrowers (up to availability of guarantees). ### Profitability channel: specification and policy adjustments - ROA decomposition used: ROAi,t = NIIi,t + Noninterest Incomei,t − LLPi,t − Otheri,t*, where Otheri,t* comprises operating expenses and write-offs of NPLs. - Dividend payout ratio assumptions: - d2020 = 0. - d2021 = min(0.2 × ∑t 2020 ROAi,t−1 Ai,t−1, 0.0015 × CET1i,t). - Moratoria modeling: - NIIi,t adjusted with term m_12 (share of accrued but non-paid interest income for m months) and country-average usage rate of moratoria θtM, and bank-specific shares of household and corporate loans sH and sC (equation A2.4). - LLPi,t adjusted for moratoria under IFRS-9 transition considerations and implied provisioning increase ΔLLPtIFRS9, with historical coverage ωt−1 and changes in corporate and household loans ΔLi,tC and ΔLi,tH (equation A2.5). - Public guarantee effects: - Guarantees allocated pro rata to banks’ share of corporate lending; governments cover losses up to φ percent of guaranteed loans; projected usage rate θtG enters provisioning reduction (equation A2.6). ### Asset and risk channels: treatment of corporate shock and guarantees - Corporate shock: single-factor surge of corporate defaults concentrated in “highly-affected” sectors mapped to banks’ sectoral corporate exposures (EBA Transparency Exercise data used where available). - For non-EU/EEA banks outside EBA coverage, sectoral exposures proxied by asset-weighted average exposures of neighboring countries as described. - Risk-weights: higher credit risk-weights applied to account for rising default risk; mitigated for newly guaranteed loans up to guarantee availability. ### Estimation approach and sample - Sample: 3,421 banks in 41 European countries (consolidated reporting, 2008–19) for ROA estimation. - Data sources: FitchConnect (bank-specific) combined with WEO macroeconomic data. - Main ROA drivers: real GDP growth yt, unemployment rate URt, lagged NPL ratio NPLi,t−1. - Panel regression specification (equation A2.8) includes bank and year fixed effects; standard errors clustered by country*year. - Stress-test horizon: t ∈ {2020, 2021}. ### Key coefficient estimates (preserved exactly as reported) - For largest euro area banks (Model 1): β1 = 0.27, β2 = −0.21, β3 = −0.02. - Annex Table 2.1—Selected coefficients (Dependent variable: Return on Assets; Model (1) SSM; Model (6) All Banks): - Real GDP growth: 0.270*** (Model 1); 0.0752** (Model 6). - Unemployment rate: −0.208** (Model 1); −0.0557 (Model 6). - NPL ratio (−1): −0.0155 (Model 1); −0.0295* (Model 6). - Observations: 4,252 (Model 1); 6,029 (Model 6). - R-squared: 0.635 (Model 1); 0.437 (Model 6). - Annex Table 2.2—NPL ratio model (Dependent variable: NPL Ratio; Model (4) All Banks): - Real GDP growth: 0.698***. - Unemployment rate: −0.0308. - NPL ratio (−1): 0.431***. - Observations: 6,186. - R-squared: 0.816. - Annex Table 2.3—Satellite models for ROA components (SSM banks): - Net Interest Income: real GDP growth 0.00656; unemployment rate 0.00535; total assets (log) (−1) −0.666***. - Non-interest Income: real GDP growth 0.0273*; unemployment rate 0.0254*; total assets (log) (−1) −0.236*. - Loan Loss Provisions: real GDP growth −0.209***; unemployment rate 0.198***; total assets (log) (−1) 0.882*. - Observations: 425 for each satellite regression. - R-squared: 0.952 (Net Interest Income); 0.763 (Non-interest Income); 0.674 (Loan Loss Provisions). ### Projection mechanics and implementation - ROA projection for 2020 (equation A2.9): ROAi,2020|Bi,2019 = ROAi,2019 + β1ROA Δy2020 + β2ROA ΔUR2020 + β3ROA ΔNPLi,2019. - NPL projection for 2020 via satellite model (equation A2.10): NPLi,2020|Bi,2019 = NPLi,2019 + β1NPL Δy2020 + β2NPL ΔUR2020 + β3NPL ΔNPLi,2019. - ROA for 2021 uses updated NPLi,2020|Bi,2019 in equation A2.11. - Satellite estimation of LLP and other ROA components (equations A2.12–A2.15) used to derive adjusted ROA with and without policy measures: - ROAi,2020 = ROAi,2020|Bi,2019 + (NIIi,2020 − NIIi,2020) − (LLPi,2020 − LLPi,2020) (equation A2.14). - ROAi,2021 = ROAi,2021|Bi,2020 + (NIIi,2021 − NIIi,2021) − (LLPi,2021 − LLPi,2021) (equation A2.15). ### Notable quantified finding (from method disclosure) - Under the baseline, the amount of NPLs of euro area banks could increase to more than €900 billion by 2021 (up from about €500 billion in 2019), which would increase the NPL ratio to more than 6 percent. *Source: Annex Figure 1.15 and Annex 2 (Methodology) content as provided in the source PDF.* ### Annex 2. Methodology ### Annex 2. Methodology ### Loan loss provisions and corporate matrix - Firms are classified into four categories by financial status: solvent-liquid (green), insolvent-liquid (pink), insolvent-illiquid (red), and solvent-illiquid (yellow). - Provisions increase for borrowers that become insolvent-liquid or solvent-illiquid. - Incidence of insolvent-illiquid borrowers determines scale and timing of write-offs, which release provisions up to the average coverage ratio. - Specification without policies (equation (A2.16)): - LLP_i,t = (1 + ρ_liquid,insolvent^C + ρ_illiquid,solvent^C − ρ_illiquid,insolvent^C)_{t−1}^{ν_t} L_i,t^C _L_i,t (β1^LLP Δy_t + β2^LLP ΔUR_t + γ1^LLP loans_to_asset_{i,t−1}) - Specification with policies (equation (A2.17)): - LLP_i,t = (1 + ρ_liquid,insolvent^C[with policy] + ρ_illiquid,solvent^C[with policy] − ρ_illiquid,insolvent^C[with policy])_{t−1}^{ν_t} L_i,t^C _L_i,t (β1^LLP Δy_t + β2^LLP ΔUR_t + γ1^LLP loans_to_asset_{i,t−1}) − (LLP_t^IFRS9 − γ_{t−1})(ΔL_i,t^C + ΔL_i,t^H)(m_12 3 ρ_t 3 = S_i,t^H S_i,t^C) − (φ − γ_{t−1}) max(μ_i,t^C ΔL_i,t^C, θ_t^G L_i,t^C) - The three relevant corporate-category terms are implemented sector-by-sector: - ρ_liquid,insolvent^C L_i,t^C = Σ_{s∈S} S_s ρ_liquid,insolvent^{C,s} L_i,t^{C,s} (A2.18) - ρ_illiquid,solvent^C L_i,t^C = Σ_{s∈S} S_s ρ_illiquid,solvent^{C,s} L_i,t^{C,s} (A2.19) - ρ_illiquid,insolvent^C L_i,t^C = Σ_{s∈S} S_s ρ_illiquid,insolvent^{C,s} L_i,t^{C,s} (A2.20) - ν_t = min(9_m ; 1) indicates share of write-offs that occur in this period (remainder 1 − ν_t in subsequent period) for deferred bankruptcies of m months since end-March 2020 within the two-year stress-test horizon. ### Profitability channel and ROA with/without policies - ROA with effective policy measures in 2020 (equation (A2.21)): - ROÂ_i,2020 = ROÃ_i,2020 | B_i,2019 3 A_i,2019_Ã_i,2020 + (NIÎ_i,2020 − NIĨ_i,2020) − (LLP̂_i,2020 − LLP̃_i,2020) - ROA with effective policy measures in 2021 (equation (A2.22)): - ROÂ_i,2021 = ROÃ_i,2021 | B_i,2020 3 Ã_i,2020_Ã_i,2021 + (NIÎ_i,2021 − NIĨ_i,2021) − (LLP̂_i,2021 − LLP̃_i,2021) ### Asset channel: lending, assets, and corporate write-offs - Credit supply equation for total loans (equation (A2.23)): - ΔL_i,t _ L_i,t−1 = a0^L + β1^L CET1_{i,t−1} + β2^L NPL1_{i,t−1} + Bank FE + ϵ_{i,t}^L - Predicted loan growth only applied if CET1_i,t ≥ MDA threshold: - MDA thresholds: 0.091 (9.1 percent) for average across euro area banks; 0.106 (10.6 percent) for non-euro area European banks. - Projected change of total assets (equation (A2.24)): - Â_i,t = (1 − d) L_i,t−1 + ΔL_i,t if CET1_i,t ≥ 0.091 - Estimated loan growth is uniform across asset classes: - ΔL_i,t / L_i,t−1 = ΔL_i,t^C / L_i,t^C = ΔL_i,t^H / L_i,t^H - δ is the average amortization rate derived as reciprocal of weighted average of uniform maturities M_C and M_H. - Corporate shock augmenting asset projection (equation (A2.25)): - Ã_i,t = (1 − d) L_i,t−1 − L_i,t−1^C ρ_illiquid,insolvent^C LGD^C (1 − ω_i) + ΔL_i,t - ρ_illiquid,insolvent^C = expected increase of debt-weighted share of illiquid and insolvent corporate borrowers relative to pre-crisis situation. - ω reflects average provisioning coverage ratio. - LGD^C is average country-specific loss given default for corporate loans (from EBA Risk Dashboard, EBA 2021a). - Write-offs added to ΔNPL_i,2020 when estimating profitability (equation (A2.26)): - NPL̃_i,2020 | B_i,2019 = NPL_i,2019 + β1^NPL Δy_2020 + β2^NPL ΔUR_2020 + β3^NPL ΔNPL_i,2019 + L_i,t−1^C ρ_illiquid,insolvent^C LGD^C (1 − ω_i) + ΔL_i,t - Consequent ROA estimate for end-2021 (equation (A2.27)): - ROÂ_i,2021 = ROÃ_i,2020 + β1^ROA Δy_2021 + β2^ROA ΔUR_2021 + β3^ROA NPL̃_i,2020 | B_i,2019 ### Policy measures, guarantees, and timing of write-offs - With policy measures, public sector guarantees affect loan growth and scale/timing of corporate defaults. - Revised asset projection with policy (equation (A2.28)): - Ã_i,t = (1 − d) L_i,t−1 − L_i,t−1^C ρ_illiquid,insolvent^C[with policy] LGD^C (1 − ω_i)(1 + φ_t^G) + ΔL_i,t + ΔL_i,t^C + max(μ_i,t^C ΔL_i,t^C, θ_t^G L_i,t^C) - ρ_illiquid,insolvent^C[with policy] recognizes mitigating impact of borrower support. - φ is uniform loss coverage by public sector proportionate to bank’s share of corporate lending in country. - Credit demand influences new corporate lending: larger of share of solvent firms μ_i,t^C and projected usage rate of guarantees θ_t^G relative to L_i,t^C. - Split of write-offs between 2020 and 2021 according to insolvency stay duration: - Asset at t = 2020 with delayed write-offs (equation (A2.29)): - Ã_i,t = (1 − d) L_i,t−1 − L_i,t−1^C ρ_illiquid,insolvent^C[with policy] LGD^C (1 − ω_i)(1 + φ_t^G) ν_t + ΔL_i,t + ΔL_i,t^C + max(μ_i,t^C ΔL_i,t^C, θ_t^G L_i,t^C) - ν_t = min(9_m ; 1) indicates share of write-offs in this period (remainder 1 − ν_t in subsequent period). - ROA with effective policy measures (equation (A2.30)): - ROÂ_i,2021 = ROÃ_i,2020 + β1^ROA Δy_2021 + β2^ROA ΔUR_2021 + β3^ROA NPL̂_i,2020 | B_i,2019 + (NIÎ_i,2021 − NIĨ_i,2021) − (LLP̂_i,2021 − LLP̃_i,2021) - NPL with policies at end of first period (equation (A2.31)): - NPL̂_i,2020 | B_i,2019 = NPL_i,2019 + β1^NPL Δy_2020 + β2^NPL ΔUR_2020 + β3^NPL ΔNPL_i,2019 + L_i,t−1^C ρ_illiquid,insolvent^C[with policy] LGD^C (1 − ω_i)(1 + φ_t^G) ν_t + ΔL_i,t ### Asset risk channel and IRB capital specification - Risk weights (RWs) measure asset riskiness; for a bank at minimum CAR, RW of 100 percent implies 8 percent of nominal exposures covered by total capital on average. - Capital requirement K based on asymptotic single risk factor IRB model (equation (A2.32)): - K = (LGD 3 [Φ^{-1}(PD) + √(R/(1−R)) Φ^{-1}(a)] − PD × LGD) × 1 / (1 + (M − 2.5)b / (1 − 1.5b)) - Statistical confidence a = 0.999 (that is, 99.9 percent). - Effective maturity M = Σ_t t × CF_t / Σ_t CF_t. - Maturity adjustment b = (0.11852 − 0.05478 × ln(PD))^2. - Correlation factor R = 0.12 × [1 − exp(−50 PD)] / [1 − exp(−50)] + 0.24 × {1 − [1 − exp(−50 PD)] / [1 − exp(−50)]}. - Asset value dynamics: dA = μA dt + σA dx (A2.34); ln A(t) = ln A(0) + μt − σ^2/2 t + σ √t X(t) with X~Φ(0,1) (A2.35). - Factor decomposition: X(t) = √R Y(t) + √(1 − R) Z(t) where Y, Z ~ Φ(0,1) independent (A2.36). - Stressed default rate at 99.9 percent confidence derived per (A2.37). ### Implicit risk-weight derivation and calibration - For minimum CAR = 8.0 percent, leverage implied = 12.5. - Implicit RW from IRB formula (equation (A2.38)): - RW = (LGD 3 [Φ^{-1}(PD) + √(R/(1−R)) Φ^{-1}(a)] − PD × LGD) / (0.08 M) × 12.5 - Derive bank-specific one-year PDs implied by observed RWs for corporate and mortgage loans (equation (A2.39)): - RW_{i,t−1}^{C;H} = (LGD_{i}^{C;H} 3 [Φ^{-1}(PD_{i,t−1}^{C;H}) + √(R/(1−R)) Φ^{-1}(a)] − PD_{i,t−1}^{C;H} LGD_{i}^{C;H}) / (0.08 M^{C;H}) - Assumed average maturities: M_C = 6.3 [years] and M_H = 22.0 [years]. - LGD_i^{C} and LGD_i^{H} vary across countries; LGDs derived as lower of net recovery rate (EBA benchmarking) and EBA Risk Dashboard LGDs. - PD estimates cross-validated with EBA general PDs and increased by up to 50 percent where EBA-reported values were significantly higher. - Pre-stress provisions for corporate and mortgage loans: LLP = (1.662 + 0.00092 (RW×100)^2 − 0.06 (RW×100)) × LGD (Jobst and Weber 2016). ### Shocking PDs and risk weights under stress and corporate shock - Change in PD consistent with change in LLP after accounting for ΔNPL (equation (A2.40)): - PD̂_{i,t}^{C;H} = PD_{i,t−1}^{C;H} × (1 + max((LLP̃_{i,t}^{C;H} − ΔNPL_{i,t}^{C;H}) / LLP_{i,t−1}^{C;H} / NPL_{i,t−1}^{C;H}, 0)) - Plug updated PDs into IRB to derive shocked RWs (equation (A2.41)): - RW̃_{i,t}^{C;H} = function(LGD_{i}^{C;H}, PD̂_{i,t}^{C;H}, R, a, M^{C;H}) per IRB formula. - Accounting for corporate write-offs in PD update (equation (A2.42)): - PD̂_{i,t}^{C;H} = PD_{i,t−1}^{C;H} × {1 / [1 + max((LLP̃_{i,t}^{C;H} − (NPL̃_{i,t}^{C;H} + L_{i,t−1}^{C} ρ_illiquid,insolvent^C LGD^C)), 0) / LLP_{i,t−1}^{C;H} / NPL_{i,t−1}^{C;H}^2]} - With policy measures, replace LLP inputs with policy-adjusted specification from profitability channel and asset-channel corporate shock (equations (A2.43) and (A2.44)): - For t = 2020 (A2.43): PD̂_{i,t}^{C;H} = PD_{i,t−1}^{C;H} × (1 + max((LLP̂_{i,t}^{C;H} − (NPL̂_{i,t}^{C;H} + L_{i,t−1}^{C} ρ_illiquid,insolvent^C[with policy] LGD^C (1 − ω_i)(1 − φ_t^G) ν_t)) / LLP_{i,t−1}^{C;H} / NPL_{i,t−1}^{C;H}, 0)) - For t + 1 = 2021 (A2.44): PD̂_{i,t+1}^{C;H} = PD_{i,t}^{C;H} × (1 + max((LLP̂_{i,t+1}^{C;H} − (NPL̂_{i,t+1}^{C;H} + L_{i,t}^{C} ρ_illiquid,insolvent^C[with policy] LGD^C (1 − ω_i)(1 − φ_t^G)(1 − ν_t))) / LLP_{i,t}^{C;H} / NPL_{i,t}^{C;H}, 0)) - Note: net impact of write-offs on portfolio credit risk weight ignored; empirical evidence suggests risk weights of defaulted loans tend to be 2.5 times higher than average loan, implying PD̃_{i,t}^{C;H} in (A2.43) would be slightly lower all else equal. ### Satellite models and estimation notes - Lending growth satellite model (equation (A2.23)) controls for lagged CET1, lagged NPL ratio, and previous loan growth; demand captured by country-year fixed effect. - Predicted loan growth applied only if CET1 ≥ MDA thresholds (0.091 euro area average; 0.106 non-euro). - Annex Table 2.4 provides estimation results for satellite models for loan growth (observations and R-squared reported in table). - Pre-stress provisioning rule-of-thumb: LLP = (1.662 + 0.00092 (RW×100)^2 − 0.06 (RW×100)) × LGD. *Source: Annex 2. Methodology, chwebfea - Annex 2. Methodology* ### Annex 2. Methodology ### Annex 2. Methodology ### Capital-to-asset ratio and leverage considerations - The capital-to-asset ratio is used as a simpler proxy for the “leverage” ratio and does not capture risk-weighted assets. - EU banks will need to maintain a minimum leverage ratio of 3 percent from June 2021. - The leverage ratio is described as a backstop to regulatory risk-weighted capital ratios (such as CET1 ratio) to limit balance-sheet swelling from risk-weight modeling uncertainties and errors. - The leverage ratio creates a simpler connection between banks’ profitability and capital accumulation. ### Key projections and stress-test methodology (profitability and asset channels) - Methodology: the exercise uses the Annex 2 methodology for the profitability and the asset channels only. - Projection: the capital-to-asset ratio is projected to fall by about 2 percentage points through 2021 under the baseline scenario with policies. - Starting level: capital-to-asset ratio at 6.9 percent in 2019. - Drivers of the decline: - European banks, on average, lose retained earnings with lower ROA through 2020 and are not able to replenish capital buffers through retained earnings even with the recovery in GDP growth envisaged for 2021 even with policies. - Higher NPLs in 2020 weigh on ROA in 2021. - Corporate bankruptcies require write-offs reducing capital against existing and projected provisions. - Policy effects: - While policies help, the capital-to-asset ratio does not benefit from reduced risk weights against a no-policy counterfactual in this exercise. - Sensitivity analysis indicates that continued dividend retention would mitigate the projected shock by 0.2 percentage points over the two-year time horizon. - Excluding the single-factor shock from a surge of corporate bankruptcies would not significantly change results given delayed insolvency procedures; most of the rising provisioning expenses are absorbed by higher profitability during the recovery in 2021. ### Stress-test channels, drivers, and scope - The solvency stress test analysis covers two main channels affecting capital and assets under stress: - Changes in profitability (net interest income and provisions). - Effective changes in total assets (net lending and write-offs after reserves). - Crisis-specific risk drivers include: 1. Write-offs due to projected insolvency of illiquid and insolvent firms (weighted by outstanding debt and mapped to sector-by-sector corporate exposure of sample banks). 2. Profitability impact of policy measures: - Lower provisions for guaranteed loans to solvent firms. - Loss forbearance on eligible loans under moratoria. - Decline in interest income due to duration of debt moratoria. - In addition, a general change in net operating income after general provisions and losses on other noninterest income due to lower GDP growth and higher unemployment rate, including impairment charges for noncorporate exposures. - The calculation does not consider changes in unexpected losses, which are reflected in the risk-weighting of asset exposures in the computation of capital adequacy. ### Accounting versus prudential provisions and IFRS 9 treatment - Accounting provisions: - Focus on proper measurement of asset values and net worth for audited, standardized, published financial statements. - Seek to reduce net value of loans toward market value; flow through the income statement and appear as negative entries on the asset side of the balance sheet. - In the EU, applicable accounting standard is transitioning from IAS 39 (incurred loss) to IFRS 9 (expected loss), with a phase-in through the end of 2022. - Accounting provisions must not fall short of required amounts nor exceed them. - Prudential provisions: - Add an extra layer of protection required by regulators and supervisors; in practice involve deductions from qualifying regulatory capital and require banks to add microprudential overlays of capital and reserves. - Some prudential provisioning requirements apply to all banks (Pillar 1); others are set bank-by-bank (Pillar 2). - In the EU, prudential provisions are specified by the Capital Requirements Regulation (CRR) and Directive (as transposed), EBA technical standards, and EBA and ECB guidelines. - IFRS 9 provisioning stages: - At origination: provision must cover expected loss resulting from possible default within 12 months (Stage 1). - If a material increase in credit risk: provision must cover lifetime expected loss on the loan (Stage 2). - If loan is impaired (usually more than 90 days past due): provision must add coverage of future accrued interest at amortized cost (Stage 3). - CRR and forbearance: - CRR specifies what measures constitute forbearance, which may trigger reclassification of the loan to nonperforming and result in higher prudential provisions. - Loan classification criteria further elaborated in EBA and ECB guidelines (2016 and 2017), with the EBA stipulating more than 90 days past due as the default threshold. - An ECB addendum on prudential provisioning issued in 2018 prompted clarifications and adjustments regarding Pillar 2 supervisory powers and Pillar 1 outcomes. *Source: Annex 2. Methodology, chwebfea - Annex 2. Methodology* --- _Source: https://www.imf.org/-/media/files/publications/dp/2021/english/chwebfea.pdf_