## 1autea2026007 ## Source details **Canonical URL:** [1autea2026007](https://www.imf.org/-/media/files/publications/cr/2026/english/1autea2026007.pdf) ## Other formats - [Markdown version](/-/media/files/publications/cr/2026/english/1autea2026007.pdf.md) - [Structured JSON version](/-/media/files/publications/cr/2026/english/1autea2026007.pdf.json) --- ### EXECUTIVE SUMMARY — Key FSAP Recommendations - Include IPS channels in solvency and liquidity stress tests and contagion analysis, and integrate liquidity channels in the contagion analysis — Agency: OeNB — Timing: ST - Broaden the use of the OeNB stress testing framework for scenario analyses beyond classical stress tests, including for macroprudential and sector-wide analyses and impact assessments — Agency: OeNB — Timing: ST - Enhance risk monitoring of banks’ large exposures to NBFIs — Agency: OeNB — Timing: MT - Enhance monitoring of CESEE large exposures and further improve surveillance and reporting — Agency: OeNB, FMA — Timing: ST - Timing legend: I Immediate (within 1 year); ST Short Term (within 1-2 years); MT Medium Term (within 3-5 years) ### Macro-financial and Banking System Findings - Austrian financial sector resilient despite domestic and international headwinds, including the ongoing Middle East conflict. - Economy experienced a broad-based recession in 2023 and 2024. - Aggregate Nonperforming Loan (NPL) pressures noted, particularly in commercial real estate (CRE) until recently. - Geopolitical risks and tighter monetary policy have constrained demand. - Consolidated banking sector assets: 255 percent of GDP as of end-2025. - Assets under management: - Investment funds: 50 percent of GDP at end-2025. - Insurance companies: 27 percent of GDP at end-2025. ### Capitalization and Profitability - Aggregate Common Equity Tier 1 capital ratio (CET1): 18.4 percent in September 2025 (from 17.9 percent at end-2024). - CET1 median: 17.6 percent for SIs and 21.8 percent for LSIs (end-September 2025). - Tier 1 capital ratio: increased from 9 percent in 2008 to more than 19 percent in September 2025. - Profitability: - Improvement driven by net interest margins from foreign activities, notably CESEE exposures. - CESEE operations comprised about a quarter of assets and around 40 percent of total profits in 2025. - About 65 percent of SIs’ profits are generated abroad; LSIs: 31 percent abroad. ### Liquidity and Funding Structure - Asset-weighted Liquidity Coverage Ratio (LCR): 185 percent in September 2025. - Net Stable Funding Ratio (NSFR): 134.8 percent in September 2025. - Household deposits: 34.7 percent of liabilities at end-2025. - Debt securities: 16.4 percent of funding. - Funding from NBFIs: 5.7 percent of funding. - Loans to NBFIs: 4.9 percent of assets at end-2025. - Household deposits represent 57 percent of LSI liabilities and 47 percent of SI liabilities (unweighted). - Currency-specific weighted average LCR: - EUR: 183 percent - CZK: 157 percent - LSIs average LCR: 240 percent; SIs average LCR: 162 percent. - 34 LSIs show LCR below 100 percent; these participate in regional liquidity associations subject to liquidity waivers. ### Banking System Structure and Interlinkages - Three main bank segments: (i) Sparkassen forming Erste Bank Group AG; (ii) Raiffeisen decentralized segment; (iii) Volksbanken segment. - At highest consolidated level: 310 credit institutions (CRR definition), 7 SIs, and 303 LSIs at end-2025. - Raiffeisen decentralized sector: three-layer structure with primary banks, eight regional Landesbanken, and RBI as regional central institution; Landesbanken hold around 60 percent of RBI (“inverse ownership”). - Institutional Protection Schemes (IPS): Raiffeisen IPS (ÖRS) covers RBI, eight Landesbanken, 265 local primary banks; Sparkassen IPS (s-Haftung) covers Erste Group Bank AG and 48 savings banks. - IPS ex-ante contributions: 0.5 percent of each bank’s RWAs; all regional and federal ex-ante funds currently fully funded. ### Key Risks and Vulnerabilities - External risks: ongoing Middle East conflict, subdued economic activity, CRE exposure, CESEE shocks. - Global risks: further geopolitical disruptions, commodity and trade risks, financial stability and sovereign stress. - Domestic risks: weak consumer and producer sentiment, potential further rise in NPLs on CRE loans. - Aggregate NPL ratio: 3 percent at end-2025. - NPL ratio on CRE loans: 8.3 percent as of end-2025. - Share of variable rate loans in new loans (H1 2025): - Households: 39 percent. - Companies: 77 percent. - Share of FX loans in total loans: declined from about ¼ at GFC time to about 9 percent. ### Stress Testing: Scenarios and Methodology (Top-down) - Coverage: 313 banks in liquidity exercises; solvency stress tests included 310 banks (data cut-off 2025:Q3). - Solvency stress tests: top-down balance sheet approach; three scenarios (baseline and two adverse) over 2026–2028 under static balance sheet assumption. - Baseline: derived from IMF Spring 2026 WEO (cut-off point February 25th 2026); baseline does not include revisions from the 2026 Middle East war. - Two updated Euro Area FSAP adverse scenarios: - Recessionary scenario: structural negative productivity shock, increased uncertainty, zero lower bound, lower inflation, higher unemployment, fiscal consolidation for high-debt EA countries. - Geopolitical scenario: rise in inflation expectations, higher commodity price volatility, fiscal support where fiscal space exists. - CESEE adjustments: more severe negative demand shocks in CESEE, shifted by one year; BVARs for Hungary, Slovakia, Romania, Croatia. - Risk types covered: credit risk, interest rate risk (IRRBB), market risk, participations risk; sensitivity analyses included. - Key modelling elements: PD models for Austria, Germany, CESEE; micro-simulation for households (HFCS 2021); LGD modelling linking LGDs to PDs and collateral effects; NII modelling via repricing ladders and pass-through econometrics. ### Solvency Stress Test Results — Aggregate and Drivers - Aggregate capital ratio remains above minimum capital requirements under severe but plausible adverse scenarios. - Decline in system-level CET1 ratio: - Recession scenario: decline of 465 basis points. - Geopolitical scenario: decline of 391 basis points. - Main contributors to capital depletion (relative to baseline): - Credit risk: 420 basis points (geopolitical); 404 basis points (recession). - Decline in NII: 171 basis points (geopolitical); 329 basis points (recession). - Market risk of SIs: 138 basis points (geopolitical); 146 basis points (recession). - Timing: most contributions occur upfront in 2026 due to expected-loss approach and accumulation of defaulted exposures. - SIs vs LSIs: - Decline for SIs: 424 basis points (geopolitical); 500 basis points (recession). - Decline for LSIs: 326 basis points (geopolitical); 396 basis points (recession). - Breach statistics: - Geopolitical scenario: 3 banks below TSCR; market share of undercapitalized banks = 0.36%; 12 banks dip into additional buffers representing 9.5 percent of assets; weighted average gap for these banks = 1.4% of RWAs. - Recession scenario: 1 bank below TSCR (less than 0.2% of assets); 12 banks dip into buffers representing 9.1 percent of assets; weighted average gap = 2.2% of RWAs. ### Market Risk and One-off Shocks - One-off market shocks applied to SIs and UniCredit’s subsidiary in Austria at start of first year and carried forward without reversal. - Example preserved metric for Austria: negative shock to real estate prices = "-32.9 percent" at the beginning of the first year. ### Net Interest Income (NII) Dynamics - Recession scenario: compression of NII in first year and thereafter driven by deposit-rate stickiness and weaker pass-through. - Geopolitical scenario: expansion of NII in first year (2026) then compression in subsequent years. - Deposit rate cumulative changes (examples from Table 5 — percentage points): - Recession Scenario cumulative over 3 years: Term deposits retail decline by 1.2 percent; sight deposits decline by 0.3 percent. - Lending rate annual shocks (selected from Table 6 — percentage points): - Consumer Loans / Mortgages (Floating Rate), Recession: 2026 -0.7; 2027 -0.9; 2028 -0.2. - Consumer Loans / Mortgages (Fixed Rate), Geopolitical: 2026 1.5; 2027 -0.3; 2028 -1.2. ### Liquidity Stress Testing and IPS Effects - LCR metrics and stress tests: - Weighted average LCR (Sept 2025): 185 percent. - Weighted average LCR under adverse scenarios: - Scenario 1: 149 percent (weighted average); SIs 126 percent; LSIs 203 percent. - Scenario 2: 133 percent (weighted average); SIs 109 percent; LSIs 190 percent. - Under Scenario 2 several SIs fall below 100 percent and total asset share of banks below 100 percent rises to 35 percent. - NSFR: - Weighted average NSFR: 140 percent for SIs and 132 percent for LSIs. - Basel III requirement / EU/EEA benchmark: 127 percent. - Cash-flow stress tests: - General risk aversion scenario: aggregate net liquidity position positive across buckets; market share of banks with shortfalls = 10 percent at 12 months (LSIs concentrated). - Idiosyncratic bank run scenario: system positive up to three months then negative; market share of banks with shortfalls = 81 percent at 12 months driven by several large SIs. - CZK net position becomes negative after three months in idiosyncratic stress; by 12 months negative net liquidity position equals 27 percent of total banking sector assets (absolute terms). - Second-round effects: IPS ex-ante funds and liquidity associations materially enhance resilience by redistributing liquidity and recapitalizing member banks in simulations. - LCR stress test with IPS modeling: Scenario 1 total excess liquidity = 7.0 percent of sector RWA; Scenario 2 = 4.2 percent. - In Scenario 1 Stage 1 (intra-group reallocation) restores all member banks below 100% using 13.4 percent of excess liquidity. - In Scenario 2 Stages 1+2 restore all member banks’ LCR consuming 74.6 percent of excess liquidity. - Forced-sale solvency sensitivity (SIs only): - Term premium shocks used (selected values in percentage points): AT 1.46; FR 2.28; DE 1.42; PL 4.94; US 2.29; GB 2.35. - Solvency impact from forced sell-off of AC government bonds: - General Risk Aversion: average 1.0% of RWA (2 weeks), 1.3% of RWA (30 days); in percent of CET1: 5.6% (2 weeks), 7.7% (30 days). - Idiosyncratic run: average 2.4% of RWA (2 weeks), 3.9% of RWA (30 days); in percent of CET1: 13.7% (2 weeks), 22.2% (30 days). - Memo: Aggregate cumulative net outflows (bn euros): GRA 23 (2 weeks) and 35 (30 days); Idiosyncratic 75 (2 weeks) and 109 (30 days). ### Contagion, Concentration, and Large Exposures - Interbank exposures are highly concentrated; contagion risks concentrated in a few larger SIs but the system is resilient to hypothetical defaults. - Contagion analysis: - Most banks neither contagious nor vulnerable. - The most contagious bank’s hypothetical default could induce system-wide loss equal to total capital of system — about four times the loss from the second most contagious bank and about ten times the average of the top twenty most contagious banks. - CESEE large exposures of SIs: 18.5 percent of their assets (Grand Total, percentage of total assets, end-December 2024). - Country-level largest exposures (percent of total assets): Czech Republic 7.2; Hungary 3.0; Romania 2.7; Slovakia 2.0; Croatia 1.5; Poland 0.9; Serbia 0.5; Slovenia 0.2; Ukraine 0.2; Albania 0.1; Bosnia and Herzegovina 0.1; Grand Total (all CESEE) 18.5. - Aggregate sector totals: Central banks 4.7; Credit institutions 1.1; Financial corporations other than credit institutions 0.3; General governments 9.6; Non-financial corporations 2.8; Grand Total 18.5. - Large exposures to NBFIs (aggregate large banks): - Before CRM: about 100 percent of CET1. - After CRM: 30 percent of aggregate CET1. - Top ten largest exposures to individual clients/groups reach more than six times aggregate CET1 before CRM; after CRM top ten reach 10 percent of aggregate CET1. ### Policy-relevant Findings and Implications - Existing macroprudential measures (KIM-V and lending guidelines) effective: borrower-based measures limiting LTVs at origination to 90 percent reduce credit risk over time. - Counterfactual borrower-based analysis (LTV<90 percent) — household annual default rates (selected): - No LTV limits, Baseline 2025: 1.27%; Geopolitical 2027: 1.79%; Recession 2027: 2.18%. - Only LTV at origination < 90 percent, Baseline 2025: 0.73%; Geopolitical 2027: 1.01%; Recession 2027: 1.22%. - Aggregate projections — difference in share of cumulative defaulted exposures (no LTV limits minus 90 percent LTV limit) (percent): - Baseline 2030: 3.02% - Geopolitical 2030: 4.03% - Recession 2030: 4.64% - Cumulative 5-year difference in defaulted exposures between no-LTV and LTV<90%: - Baseline: about 3 percentage points ≈ €5.6 billion reduction. - Geopolitical: about 4–4.6 percent ≈ €7.5 billion / €8.7 billion reduction (source lists €7.5 billion and €8.7 billion for geopolitical twice). - Policy recommendations: - Continue enhanced monitoring of CESEE large exposures and NBFI exposures. - Continue oversight of IPS and liquidity associations given their role in dampening vulnerabilities and potential to amplify interconnections under stress. - Central bank liquidity provision is important during market stress to avoid forced sell-offs that would significantly impact solvency. ### Scope, Data Vintage, and Limitations - Data cut-off: supervisory data up to 2025:Q3; balance sheets and P&L validated and cleaned by OeNB as of end-2024 and updated to 2025:Q3. - Solvency test coverage: 310 banks (solvency) and 313 banks (liquidity), representing over 95 percent of banking system total assets. - Static balance sheet assumption: portfolio composition and liability structure remain unchanged; no write-offs; new originations equal maturing loans; conservative (upper-bound) capital depletion estimates. - IPS and liquidity association modelling: four-stage mutual support framework; assumptions include instantaneous transfers and proportional allocation rules; limitations noted (operational frictions, monetization costs, timing). - Contagion analysis limited to Austria-only cascade effects; cross-border contagion not modeled. - Reverse stress tests and sensitivity analyses performed; results depend on model assumptions (e.g., fire-sale haircut δ calibrated at 35 percent). *Source: IMF staff analysis and figures from the Austria FSAP chapter (data cut-off 2025:Q3).* ### EXECUTIVE SUMMARY __________________________________________________________________________ 8 ### EXECUTIVE SUMMARY ### Key FSAP Recommendations (Table 1) - Include IPS channels in solvency and liquidity stress tests and contagion analysis, and integrate liquidity channels in the contagion analysis — Agency: OeNB — Timing: ST - Broaden the use of the OeNB stress testing framework for scenario analyses beyond classical stress tests, including for macroprudential and sector-wide analyses and impact assessments — Agency: OeNB — Timing: ST - Enhance risk monitoring of banks’ large exposures to NBFIs — Agency: OeNB — Timing: MT - Enhance monitoring of CESEE large exposures and further improve surveillance and reporting — Agency: OeNB, FMA — Timing: ST (Note: Timing legend — I Immediate (within 1 year); ST Short Term (within 1-2 years); MT Medium Term (within 3-5 years)) ### Macro-financial and Banking System Findings - The Austrian financial sector has remained resilient in recent years despite domestic and international headwinds, including the ongoing Middle East conflict. - The economy experienced a broad-based recession in 2023 and 2024. - Rising NPLs were noted, particularly in the commercial real estate segment until recently. - Geopolitical risks and tighter monetary policy have constrained demand. ### Capitalization and Profitability - Aggregate Common Equity Tier 1 capital ratio (CET1) reached 18.4 percent in September 2025, from 17.9 percent at the end of 2024. - The improvement in CET1 reflects continued strong profitability in recent years, in particular from net interest margins driven by foreign activities. - Profitability has declined since its peak in 2023 but remained relatively high with a return on equity at [value not provided in source excerpt]. *Source: EXECUTIVE SUMMARY — 1autea2026007* ### 8.9 percent in percent in 2025. Liquidity buffers are high, with weighted Liquidity Coverage Ratio ### 1autea2026007 - 8.9 percent in percent in 2025. Liquidity buffers are high, with weighted Liquidity Coverage Ratio ### Overview - Consolidated banking sector assets account for 255 percent of GDP as of end-2025. - Assets under management of investment funds and insurance companies represent respectively 50 and 27 percent of GDP at end-2025. - The asset-weighted Liquidity Coverage Ratio (LCR) was 185 percent in September 2025. - Aggregate Net Stable Funding Ratio (NFSR) was 134.8 percent in September 2025. - Less Significant Institutions (LSIs) have on aggregate higher capital ratios and higher liquidity buffers than Significant Institutions (SIs). ### Banking system structure and interlinkages - Three main bank segments: (i) Sparkassen forming Erste Bank Group AG; (ii) the Raiffeisen decentralized segment; and (iii) the Volksbanken segment. - At the highest consolidated level: 310 credit institutions based on the CRR definition, 7 SIs, and 303 LSIs at end-2025. - SIs and LSIs in the decentralized Raiffeisen sector have complex ownership structures and financial interlinkages, including layers of ownership and mutual cooperation arrangements—institutional protection scheme (IPS) of mutual support, and liquidity associations. - CESEE operations comprised about a quarter of the assets and around 40 percent of the total profits of the Austrian banking sector in 2025. - The two largest banks, Erste Group Bank and Raiffeisen Bank International (RBI), account for 44 percent of banking assets; the subsidiary of UniCredit in Austria accounts for 9 percent of banking assets. - Total foreign exposures of the banking system reached 26.2 percent of assets at the end of 2025, and 41.6 percent for larger banks (joint stock banks). - On an unconsolidated basis, loans to non-financial private sector account for about 47 percent of assets, and securities holdings are significant (data sources: OeNB website). ### Key risks and vulnerabilities - Key external risks: ongoing conflict in the Middle East, subdued economic activity, exposure to CRE lending, and shocks in the Central, Eastern and Southeastern European (CESEE) regions. - Global external risks: materialization of further geopolitical disruptions, commodity and trade risks, and financial stability and sovereign stress. - Austria-specific risks: disruptions in CESEE countries causing lower growth, which could impact Austrian banks’ profitability through credit risk and interest rate risk. - Domestic risks: persistent weak sentiment among consumers and producers could weigh on private consumption and investment demand. - A further rise in NPLs on CRE loans could have a pronounced impact on banks’ profitability. ### Stress testing results and scenarios - Solvency stress tests relied on an update of the Euro Area FSAP adverse scenarios with more severe negative demand shocks in CESEE, shifted by one year, and on the IMF’s Spring 2026 WEO data as available before the war in the Middle East for the baseline. - The aggregate capital ratio of the Austrian banking system remains above capital requirements under severe but plausible adverse macroeconomic scenarios. - Credit risks and interest rate risks are the main drivers of aggregate capital depletion in the adverse scenarios. - Capital ratio decline is smaller in the geopolitical scenario than in the recession scenario because net interest income compresses more in the recession scenario despite larger credit risk in the geopolitical scenario. - The decline in capital ratios is larger for SIs than for LSIs, reflecting larger credit risk and larger decline in net interest income including from CESEE exposures. - Under adverse scenarios: only a small number of LSIs breach their Total SREP Capital Requirement (TSCR), and several additional banks, including three SIs, moderately dip into their additional buffers. ### Liquidity resilience and IPS effects - Two stress scenarios were considered for both the LCR stress tests and the cash-flow stress tests, derived from the scenarios of the 2025 Euro Area FSAP. - The asset-weighted LCR remains above regulatory requirements under the two stress scenarios, for both SI and LSI segments, and for both total currency and significant currencies. - Cash-flow stress tests show both SIs and LSIs, on average, can withstand significant liquidity outflows by drawing down counterbalancing capacity over extended horizons exceeding 30 days, under both a generalized risk aversion scenario and an idiosyncratic bank run scenario. - Pockets of vulnerable SIs and LSIs emerge, especially beyond the 3-month horizon. - IPS ex-ante funds of the decentralized sector appear sufficient to recapitalize banks experiencing a capital shortfall to minimum requirements under the two adverse scenarios of the solvency stress tests; regional members have sufficient capital surplus on aggregate to recapitalize banks to requirements including buffers. - Liquidity stress tests incorporating second-round effects from IPS funds and liquidity associations underscore the importance of these mechanisms in preserving resilience to liquidity shocks by channeling liquidity surplus through redistribution among member banks and activation of IPS mechanisms. - The IPS of the Sparkassen is not considered in the analysis because the Sparkassen are consolidated into one banking group. ### Contagion, concentration, and large exposures - Contagion stress test suggests Austrian interbank exposures are highly concentrated, but the system is resilient to hypothetical defaults; only a limited number of banks would create contagion risks. - Baseline case: contagion risks are highly concentrated to originate from a few larger banks; the overwhelming majority of banks were found neither contagious nor vulnerable. - Only a few SIs were found contagious, but their liquidity positions and asset pools and capital prevent them from being vulnerable. - Most vulnerable LSIs with relatively tight liquidity or capital positions are small. - IPS scheme generally lowers vulnerability of participating institutions but may amplify interconnections in case of stress to larger banks; larger banks may have to contribute relatively more than their pro-rata share, potentially increasing spillover effects up to regulatory limits. - CESEE large exposure of SIs accounts for 18.5 percent of their assets. - Potential losses from CESEE would increase contagion risks as capital loss would be concentrated on banks creating cascading effects through interconnections in the interbank market. - Need for enhanced monitoring of large exposures in the CESEE region. ### Macro-financial background (selected points) - The Austrian economy experienced two years of recession in 2023–24. - Public debt rose to nearly 82 percent of GDP in 2025. - The Middle East conflict resulted in a downward revision of economic growth by 0.5 percentage points for 2026 and by 0.2 percentage points in 2027. - Household debt relative to net disposable income declined from 94 percent in 2019 to 74 percent at end-June 2025. - Overall credit growth was at 2.6 percent y-o-y in March 2026. - Mortgage lending appears to have bottomed out; corporate lending growth has been picking up slowly. ### Policy-relevant findings and implications - Existing macroprudential policy measures under the KIM-V legislation and guidelines concerning standards for mortgage origination are effective: borrower-based measures limiting LTVs at origination to 90 percent notably reduce credit risk over time. - Central bank liquidity provision is important during market stress events because a hypothetical forced sell-off of amortized cost fixed income securities would have a significant impact on solvency. - Enhanced monitoring of large exposures, particularly to CESEE countries, is warranted given their size and potential to amplify contagion. - Continued oversight of IPS and liquidity associations is critical given their role in dampening vulnerabilities and their potential to amplify interconnections in stress scenarios. *AUSTRIA INTERNATIONAL MONETARY FUND* ### 13.6 percent of assets at the end of 2025. Household deposits remain the main source of funding at ### 1autea2026007 - 13.6 percent of assets at the end of 2025. Household deposits remain the main source of funding at ### Funding structure and balance-sheet composition - Household deposits account for 34.7 percent of liabilities at the end of 2025. - Debt securities account for 16.4 percent of funding. - Funding from NBFIs accounts for 5.7 percent of funding. - Loans to NBFIs are 4.9 percent of assets at the end of 2025. - Loans to households, loans to NFCs, loans to NBFIs, loans to banks, loans to official sector, cash balances, securities, equity and participations and other assets are presented in Figure 4 as shares of total assets (end 2025) by type of bank (values shown in the source figures). ### Raiffeisen segment: decentralized structure and liquidity arrangements - The Raiffeisen sector has a three-layer structure: - First tier: local primary banks (primary banks hold shares in one of the eight regional Landesbanken). - Second tier: eight regional Landesbanken acting as central institutions to primary banks. - Third tier: RBI (Raiffeisen Bank International) as the central institution at the regional level. - The Landesbanken hold around 60 percent of RBI (“inverse ownership”). - Local banks and respective Landesbanken form local liquidity associations where Landesbanken are central institutions and primary banks upstream liquidity surplus. - Some liquidity associations have been granted a waiver on LCR requirements under EU regulations; as a result, some LSIs of the decentralized sector can legally report LCR below 100 percent because liquidity is upstreamed and managed at the liquidity associations' levels. - An Institutional Protection Scheme (IPS) covers the entire decentralized sector with IPS-internal burden-sharing mechanisms between regions. - IPS ex-ante contributions account to 0.5 percent of each bank’s RWAs. - Each member must contribute to the ex-ante fund available for support, amounting to 0.5 percent of each bank’s RWAs. - All regional and federal ex-ante funds are currently fully funded. - If the ex-ante fund is not sufficient, members are required to provide ex-post contributions. - Every member of the IPS must first exhaust all reasonable resources at the regional level before resorting to IPS rights. ### Institutional Protection Schemes (Box summary) - Austria has two IPSs (Raiffeisen IPS and Sparkassen IPS), both approved under Article 113(7) CRR and operating under ongoing supervision. - Raiffeisen IPS, managed by Österreichische Raiffeisen-Sicherungseinrichtung eGen (ÖRS), includes: - Raiffeisen Bank International AG (RBI), - Eight regional Landesbanken, - 265 local Raiffeisen primary banks. - Sparkassen IPS, managed by s-Haftungs GmbH (s-Haftung), comprises: - Erste Group Bank AG, Erste Bank der Österreichischen Sparkassen AG, and all Austrian savings banks (Sparkassen), totalling 48 institutions. - IPSs provide preventive liquidity and solvency support financed by prefunded resources and member contributions. - IPSs are recognized as deposit guarantee and investor compensation schemes (DIS); liability companies (ÖRS and S-Haftung) operate the DIS of respective IPSs. IPS and DIS funds are legally and operationally segregated. - The Raiffeisen sector’s inverse ownership (“inverted pyramid”) and decentralization eliminate consolidation requirement, though ÖRS prepares extended aggregated balance sheet for the whole IPS scheme according to Article 113(7)(e) CRR. - Prudential benefit: IPS membership provides a zero percent risk weight for qualifying intragroup exposures. - Risk trade-off: extensive mutual support and preferential prudential treatment increase financial interdependence and can enhance stability in going-concern context but give rise to material contagion risks in stress scenarios. ### Profitability, income composition, and foreign activities - Profitability has increased in recent years, driven mostly by net interest income (NII) in the higher interest rate environment, notably from exposures in CESEE countries. - SIs tend to experience higher returns on assets (ROA) than LSIs. - About 65 percent of SIs’ profits are generated abroad, compared to 31 percent for LSIs. - The recent increase in operating gains results largely from higher net interest income (NII). - Provisions increased from 2023 onwards relative to 2022. - SIs generate more than ½ of their profit abroad while LSIs generate about 2/3 domestically (source figures). ### Capital and liquidity buffers - Common Equity Tier one capital to risk weighted assets ratio (CET1 ratio) reached 18.4 percent at end-September 2025. - CET1 ratio median: 17.6 percent for SIs and 21.8 percent for LSIs. - Asset-weighted Liquidity Coverage Ratio (LCR) is 185 percent. - Net Stable Funding Ratio (NSFR) is 134.8 percent. - Tier 1 capital ratio increased from 9 percent in 2008 to more than 19 percent in September 2025. - Share of FX loans in total loans declined from about ¼ at the time of the GFC to about 9 percent. - Since the 2020 FSAP, the banking system has consolidated and improved profitability and capitalization. ### Vulnerabilities and risks - Aggregate NPL ratio (nonperforming loans to total gross loans) stood at 3 percent at end-2025. - NPL ratio on CRE loans rose to 8.3 percent as of end-2025; NPL ratio on non-CRE loans declined slightly. - Borrower-based measures (BBMs) for RRE financing were introduced in August 2022; they were allowed to expire in July 2025 and replaced by lending guidelines in line with BBM limits. - Share of variable rate loans in total new loans (H1 2025): - Households: 39 percent. - Companies: 77 percent. - Potential structural vulnerabilities: - Interconnectedness via liquidity associations and IPS could propagate shocks through the decentralized sector. - NBFI risks are relatively contained by size but merit attention due to interconnectedness with banking sector, CRE, and CESEE exposures. - Key future risks: - Global: increased global trade policy uncertainty, geopolitical tensions, commodity price shocks, continuation of Middle East conflict leading to high energy prices, higher interest rates, and heightened financial market volatility. - Domestic: subdued economic activity, exposure to CRE lending, shocks in CESEE region, weak sentiment among consumers and producers. - Manufacturing sector’s share in NPLs was 16 percent in 2024. - A further rise in NPLs on CRE loans could have a pronounced impact on banks’ profitability. ### Scope, methodology, and scenarios for the systemic risk analysis - Coverage: 313 banks included in the exercise, accounting for over 95 percent of banking system’s total assets. - Note: solvency stress tests included only 310 banks due to incomplete data for 3 banks. - Data cut-off: supervisory data up to 2025:Q3. - Solvency stress test: - Top-down balance sheet approach. - Considered three scenarios: a baseline derived from IMF economic forecasts with a cut-off point of February 25th 2026, and two adverse scenarios. - Both adverse scenarios cover a three-year horizon 2026-2028 under a static balance sheet assumption. - Included sensitivity tests and second-round analysis of IPS in decentralized sector. - Used OeNB’s ARNIE with inputs estimated from FSAP granular satellite model estimations. - Baseline does not include revisions to forecasts due to the war in the Middle East. - Liquidity stress test: - Top-down exercise using regulatory data. - Assessed liquidity risks using LCR tests and cash-flow stress tests, including second-round effects from liquidity associations and IPS. - Considered two stress scenarios from the 2025 Euro Area FSAP liquidity stress tests: - (i) a general risk aversion scenario, and - (ii) a more severe idiosyncratic bank run scenario. - Reverse stress tests were also performed. - Contagion and interconnectedness analysis: - Network analysis to assess default cascades via credit and funding/liquidity channels. - Counterfactual scenario to demonstrate the role of IPS in the decentralized sector. - Cross-border large exposures to CESEE countries assessed at a granular level with data corrections applied. *Source: IMF staff analysis and figures from the Austria FSAP chapter (data cut-off 2025:Q3).* ### 15. The solvency stress tests assessed the stability of the Austrian banks against a baseline ### The solvency stress tests assessed the stability of the Austrian banks against a baseline scenario and two updated versions of the Euro Area FSAP adverse scenarios ### Macro-financial scenarios: structure and key distinctions - Scenarios assessed: baseline and two updated versions of the 2025 Euro Area FSAP adverse scenarios (recessionary and geopolitical), updated with IMF Spring 2026 WEO data as available at the time of the main mission in March for the baseline. - Common features of both adverse scenarios: - international trade and geopolitical frictions; - lower aggregate demand due to loss of confidence; - endogenous monetary policy responses; - financial stress with higher term premia; - lower asset and house prices. - Recessionary scenario specifics: - structural negative shock to productivity growth; - increase in uncertainty together with a zero lower bound on monetary policy; - lower inflation, increase in unemployment, and lower commodity prices; - fiscal reaction: fiscal consolidation for medium- and high-debt eurozone countries. - Geopolitical scenario specifics: - rise in inflation expectations; - heightened volatility in commodity prices; - fiscal reaction: fiscal support for countries with appropriate fiscal space in the eurozone. - Relevance relative to 2026 conflict forecasts: - The adverse scenarios were defined prior to the outbreak of the 2026 Middle East conflict but "remain relevant" because they are significantly more severe than the 2026 Article IV forecasts and 2026 Spring WEO severe scenario, particularly regarding shocks to real GDP growth, unemployment and inflation. ### Additional CESEE-focused adjustments and contagion handling - Given high Austrian bank exposure to CESEE, additional macro-financial scenarios were set up to ensure consistency with euro-area and Austria settings; scenarios feature more severe negative demand shocks in CESEE, shifted by one year. - For Hungary, Slovakia, Romania and Croatia, a BVAR model was developed to generate adverse scenarios anchored to Euro Area FSAP shocks. - Contagion analysis focus: - the structure of the interbank market where interbank exposures are highly concentrated; - large interbank exposures within the decentralized sector and strong connections between other banks and large decentralized-sector banks; - the prominent role of large decentralized-sector banks in interbank liquidity flows; - features of complex protection schemes and CESEE large exposures (hypothetical loss in credits of nonfinancial corporations in CESEE assessed for contagion and vulnerability implications). ### Liquidity stress testing - Liquidity stress test scenarios derived from the 2025 Euro Area FSAP. - Additional stress scenarios (LCR and cash-flow based): - general risk aversion scenario; - idiosyncratic bank runs scenario. - Under these scenarios sovereign stress events lead to: - larger valuation haircuts on HQLA; - elevated secured and unsecured outflows. - Second-round effects assessed from IPS and liquidity associations given their strong role in the decentralized sector. ### Market risk: one-off severe market shocks and calibration - SIs and UniCredit’s subsidiary in Austria are assessed against one-off severe market shocks that materialize in the first year of each adverse scenario; the capital impact is carried over in the two following years and shocks are not reversed during the 3-year macro scenarios. - Two sets of one-off market shocks (aligned with each macro scenario) applied to SIs and UniCredit’s subsidiary: - Geopolitical-aligned market shocks: increase in interest rates (particularly at the short end of the EUR yield curve), higher commodity prices, wider credit spreads, and a drop in equity prices; include volatility shocks to equities and interest rates. - Recessionary-aligned market shocks: drop in commodity and equity prices, an increase in credit spreads of mid-risk Euro area sovereigns (Austria is low-risk sovereign) and high-risk Euro Area sovereigns, while the short end of the EUR yield curve remains stable; include volatility shocks to equities and interest rates. - Market risk add-ons calibrated to capture high-frequency market price and volatility movements using an Expected Shortfall approach (expected shortfall at 0.1 percent of market factors marginal distributions over holding periods of 20 or 60 days, as in the 2025 Euro Area FSAP). - Market scenario example metric preserved from methodology: - For Austria, negative shocks to real estate prices are applied at the beginning of the first year and there is no recovery after the first year; for Austria, the shock is "-32.9 percent" at the beginning of the first year. ### Scope, data vintage, and coverage - Entities covered: - 7 Austrian SIs; - UniCredit’s subsidiary in Austria; - more than 300 LSIs. - Coverage represents "about 95 percent of bank assets." - Data and vintage: - balance sheets and P&L validated and cleaned by the OeNB as of end-2024; - updated with 2025:Q3 data for capital ratios and interest income and expenses, and portfolio and country-specific exposures. - Supervisory data sources: ECB provided FINREP and COREP templates and ECB Short-Term Exercise (STE) files on IRRBB and market sensitivities; OeNB provided Austria credit register and MFI interest rate statistics. - OeNB guidance on intra-year data: advised against relying entirely on intra-year quarterly data due to seasonality; cumulative quarterly values between 2024:Q3 and 2025:Q3 were considered to smooth seasonal effects in P&L variables of net income. ### Stress test methodology and key assumptions - Exercise type: top-down FSAP stress test (IMF in-house models), distinct from EBA-SSM-ECB constrained bottom-up exercise. - Risk types covered: credit risk, interest rate risk, market risk and participations risk; complementary sensitivity analyses performed. - Balance sheet treatment: standard static balance sheet assumption: - portfolio composition and liability structure remain unchanged; - no write-offs (and no cures); - new originations equal maturing loans; - nominal balance sheet of each bank remains of constant size nominally; - implies defaulted exposures accumulate on banks’ balance sheets and yields conservative (upper-bound) estimates of capital ratio declines from credit risk. - Consolidation and data treatment: IRB and STA portfolios consolidated at bank level; credit risk modeling uses an expected loss approach; Initial Point-in-Time (PiT) PDs and Through the Cycle (TTC) PDs for IRB portfolios are those reported by each SI in supervisory templates; STA PDs estimated from recent historical reported transition matrices. - Capital hurdle rates: - First hurdle (bank-specific): minimum requirement = Basel III common 4.5 percent minimum plus bank-specific Pillar 2 Requirements (Total SREP Capital Requirement, TSCR). - Second hurdle: adds capital conservation buffer of 2.5 percent and systemic buffers (any O-SII buffer and the general and sectoral systemic risk buffer). - Market risk add-ons for SIs rely on modelling from the 2025 Euro Area FSAP for calibration. ### Credit risk modelling for exposures to non-financial corporations (NFCs) - Country-specific econometric models developed for PDs for each country of exposure of Austrian banks. - Austria PD model: - based on bank-sectoral panel regressions of the logistic transformation of default rates on macroeconomic indicators; - microdata source: Austrian credit register panel starting in 2008. - Germany PD model: similar bank-level panel regression approach (methodology initially developed for the 2022 Germany FSAP). - CESEE PD models: derived from country-specific time series regressions of the logistic of Expected Default Frequencies (EDFs) on macroeconomic variables. - PD modelling details for Austria and Germany (Box 2 summary): - dynamic panel regressions estimated for each NACE sector (level 1 aggregation) at bank level at quarterly frequency using Austrian credit register data cleaned by OeNB; - main specification includes lagged logistic default frequency, year-on-year real GDP growth, average EURIBOR/Euro STR (r_{t-1}), a post-pandemic dummy (post = 0 before 2021:Q4, 1 thereafter), and quarter pandemic dummies; - specific treatment for real estate related sectors (NACE F and L): pooled separately due to larger sensitivity to economic activity and interest rates; other sectors pooled together with sector fixed effects. *Source: IMF staff, solvency stress tests section of the Austria FSAP chapter (validated OeNB data as of end-2024 and updates to 2025:Q3).* ### Box 2. Austria: Probabilities of Default Models for Exposures to NFCs in Austria and ### Box 2. Austria: Probabilities of Default Models for Exposures to NFCs in Austria and Germany (continued) ### Econometric model for German exposures - Model: quarterly pooled OLS regressions of default rates on macroeconomic variables (pooled bank-sectoral level panel regression developed at the Bundesbank). - Data: default rates constructed from German credit registry data cleaned and aggregated at the bank sectoral level by the Bundesbank. - Estimation period: 2008: Q1 to 2019: Q4. - Pooled sectoral regression specification: - y_s,t = β · y_s,t−1 + α · X_t + Q_t + ε_s,t (2) - y_t is the default rate on loans (share of outstanding loans for which a default event occurs during a given quarter). - X_t is a vector of macroeconomic variables. - Q_t is a set of quarterly dummy variables to control for seasonality. - ε_t is a residual. - The econometric model relied upon is the one reported in the first column in the table (see supplied table excerpt). ### Final regressions for Austria (NFC exposures) - Regressions reported separately for "Other sectors" and "Real Estate". - Key reported coefficients and statistics (as presented): - Lagged dependent variable: 0.915*** (Other sectors); 0.896*** (Real Estate) - Euribor (1 month) (t-1): 0.00668 (Other sectors); 0.0263*** (Real Estate) - YoY Real GDP growth: -0.762*** (Other sectors); -1.520*** (Real Estate) - YoY Real GDP growth *post: -0.515 (Real Estate column only) - Constant: -0.236*** (Other sectors); -0.316*** (Real Estate) - Observations: 64,516 (Other sectors); 23,026 (Real Estate) - R2: 0.923 (Other sectors); 0.884 (Real Estate) - Robust standard errors in parentheses (cluster by quarter) - Significance: *** p<0.01, ** p<0.05, * p<0.1 - Dependent variable: logistic transformation of annualized quarterly default frequency. - Real estate: NACE sectors F & L. - Regressions include bank and NACE sectoral fixed effects and dummy indicators for the Covid period (one indicator per quarter between 2020:Q1 and 2021:Q4). - Note: "This specification is the final one, after extensive search with various indicators of economic activity and longer-term interest rates, in addition to the variables retained in equation 1." (footnote 1/) ### PD elasticities for NFCs and cross-border exposures (findings) - PD elasticities for exposures in Austria and for cross-border exposures are reported in Table 4 and Figure 32a (referenced in text). - Key findings: - For exposures in Austria: - Default rates in real estate related sectors (NACE sectors F and L) are significantly more sensitive to macro-financial fluctuations than in other sectors. - Sensitivity of real estate sectors has increased since the pandemic. - For real estate sectors, increase in default rates is higher in the geopolitical scenario (a doubling of default rates at the peak) than in the recession scenario — driven by a significant effect of shocks to short-term interest rates. - For other sectors, most evolution in default rates is driven by shocks in real GDP growth; little difference across the two scenarios. - For cross-border exposures: - Very large differences in elasticities across countries in the adverse scenarios. - Example extremes: - Germany: little increase in PDs (a 30 percent increase at the peak). - Hungary: PDs increase by more than 10 times in the geopolitical scenario (significant sensitivity to shocks to interest rate and real GDP growth). ### PD elasticities table excerpts (All sectors / non real estate / Real estate, scenarios) - Values (multiples of 2025 value) by year and scenario as reported: - Baseline Scenario: - 2025: 1 / 1 / 1 - 2026: 0.99 / 1.00 / 1.06 - 2027: 0.98 / 1.00 / 1.25 - 2028: 0.97 / 1.03 / 1.26 - Geopolitical Scenario: - 2025: 1 / 1 / 1 - 2026: 1.28 / 1.04 / 1.13 - 2027: 2.05 / 1.26 / 1.80 - 2028: 1.87 / 1.27 / 1.59 - Recession Scenario: - 2025: 1 / 1 / 1 - 2026: 1.00 / 1.00 / 1.06 - 2027: 1.25 / 1.00 / 1.25 - 2028: 1.26 / 1.03 / 1.26 ### PD modelling for CESEE (Box 3 summary) - Approach: country-by-country time series regressions of average Expected Default Frequencies (EDFs) for corporates from Moody’s KMV, starting 2000: Q1. - Specification: - logistic_EDF_t = β · logistic_EDF_t−1 + α · X_t + ε_t (3) - logistic_EDF_t is the logistic transformation of the EDF; X_t is a vector of macro-economic variables; ε_t is a residual. - Final specifications include dummy indicators for each quarter of the pandemic between 2020: Q1 and 2021: Q4. - Sample reported coefficient excerpts for selected countries: - Hungary: Lagged dependent variable 0.937***; 10 year bond yield 0.010; Quarter-on-Quarter real GDP growth -7.995**; Constant -0.365; Observations 93; R2 0.879. - Czech Republic: Lagged dep var 0.839***; 10 year bond yield 0.0328; Quarter-on-Quarter real GDP growth -6.44; Constant -1.295***; Observations 93; R2 0.863. - Slovakia: Lagged dep var 0.586***; 10 year bond yield 0.0140; Quarter-on-Quarter real GDP growth -12.56**; Constant -1.729***; Observations 93; R2 0.428. - Romania: Lagged dep var 0.529***; 10 year bond yield 0.0210; Quarter-on-Quarter real GDP growth -4.844; Constant 0.215; Observations 62; R2 0.978. - Additional panel-style results (FE / no FE) for other CESEE regressions reported with coefficients for lagged dep var, Real GDP growth, FCI, quarter dummies, constants, Observations (680) and R2 (various: 0.551, 0.512, 0.565, 0.515, 0.568, 0.520). ### Credit risk for Households: micro-simulation overview - Model: micro simulation model based on the ECB Household and Finance Consumption Survey (HFCS) 2021 to estimate PDs for households. - Rationale: micro data allow estimation of default risk accounting for household total indebtedness, debt service, gross financial assets (buffers), wage income, and remaining loan-to-value (LTV) on mortgage debt. - Assumption: mortgages are full recourse in European countries; households draw down financial assets to repay debt and default if and only if financial assets are fully depleted and other income insufficient to cover debt service and necessary spending. - Households may default only if they become unemployed. - Model estimates default probabilities for households with mortgages and for households with only consumer loans. - Benchmarking: evolution of PDs is benchmarked on starting point PDs reported in the EBA Risk Parameter Statistics as of 2025: Q3. ### Micro-simulation model key equations (Box 4) - PD decomposition: - PD_i,t = Proba[Unemployment]_i,t × conditional_PD_i,t (1) - PD_i,t is the probability of default of household i at date t. - Proba[Unemployment]_i,t is the probability of unemployment of household i at date t. - conditional_PD_i,t is the probability of default conditional on being unemployed. - conditional_PD_i,t = (1 − Proba[employment]_i,t) × [1 | Financial_Assets_i,t < 0] (2) - [1 | Assets_i,t < 0] = 1 if gross financial assets are negative, 0 otherwise. - Implies an unemployed household will default if it remains unemployed and its stock of financial assets is entirely depleted at date t. - Condition for default if unemployed: - Financial_Assets_i,t < 0 (3) - Asset dynamics if employed: - A_i,t = (1 + r) × A_i,t−1 + other_inc_i,t + Wage_i,t − debt_services_i,t − Spending_GS_i,t (4) - Asset dynamics if unemployed: - A_i,t = (1 + r) × A_i,t−1 + other_inc_i,t + UB_i,t − debt_services_i,t − Spending_GS_i,t (5) - Unemployment benefit: - UB_i,t = Net_replacement_t,r × Wage_i,t (6) - Net_replacement_t,r is the net replacement rate at date t, r quarters after becoming unemployed. - Spending truncation: Spending_GS_i obtained from the survey and truncated at the 10th percentile (lower bound) and the median (upper bound). - Replacement rates and likelihood of exit from unemployment are estimated from OECD macro data; for adverse scenarios outflows from unemployment are calibrated on the GFC period, and for baseline on the pre-pandemic period. - Note: toolkit derived from Gross, M., Tressel, T., Ding, X., and Terenau, E., 2022, “What Drives Mortgage Default Risk in Europe and the US?”, IMF Working Paper No. 2022/065 (referenced in source). ### PD paths for Households (findings) - PD paths for domestic and cross-border exposures are reported in Figures 10 and Figure 32b (referenced). - For most country exposures, PDs for mortgages more than double in the adverse scenarios relative to their starting points. - Largest increases for mortgage PDs: exposures to Germany, Czech Republic and Slovakia. - Smallest increases for mortgage PDs: Croatia. - PDs for consumer loans tend to increase moderately more than PDs for mortgages. - Note on Romania: no household survey available; Croatia micro-simulations combined with Romania macro scenarios and starting point PDs used as proxy for household exposures in Romania. ### LGDs for unsecured exposures and collateral effects - LGDs for unsecured exposures modelled linking LGDs to fluctuations of PDs and an initial bank-level value for each geographical segment (methodology from Frye and Jacobs (2012), Vašíček-type equation). - LGD formula used: - LGD_t0+h = Φ(Φ^−1(PD_t0+h) − k) / PD_t0+h - Parameter k: - k = [Φ^−1(PD_t0^*) − Φ^−1(PD_t0^* × LGD_t0^*)] / √(1 − ρ) - LGD starting point data obtained from actual recoveries from banks; method used in IMF’s Global Bank Stress Test. - Example starting aggregate LGDs as of 2025: Q3 in Austria: - NFCs: 31.6 percent - Mortgages: 20.19 percent - Consumer loans: 51.8 percent - For exposures secured by real estate and other collateral, an additional adjustment applied to unsecured LGD to account for collateral value loss from house price decline: - LGD_t = LGD_unsecured,t × (1 − (real_estate_collateral_EAD × (House_price_t / House_price_0) − Other_collateral_EAD)) ### Risk-Weighted Assets (RWAs) - Projection of RWAs accounted for: - defaulted credit exposures formation, - changes in credit quality characteristics of IRB portfolios, - rating migrations of securitization portfolios. - Assumptions: - RWA kept constant (at the level of the cut-off date) for STA portfolios. - RWAs for operational and market risk kept constant at the cut-off date. - Static balance sheet assumption implies capital impact due to RWA changes confined to deteriorating quality of IRB portfolios. ### Net Interest Income (NII) modelling - Method: semi-structural repricing gaps methodology combining econometric models at country, portfolio segment level and maturity “repricing” ladders for each bank. - Two steps: - 1) Develop econometric models for interest rates on new business for loans and deposits to quantify pass-through from market rates and macro-conditions to deposit and lending rates (Box 6). - For Austria: models developed at bank-instrument-counterparty sector using confidential Austria MFI Interest Rate Statistics (micro-data). - For Germany and CESEE: regressions at country-instrument-counterparty sector based on aggregate EU MFI Interest Rate Statistics. - 2) Use IRRBB bank-level repricing structure to estimate timing of transmission of market rate shocks to deposit and lending rates by instrument; combine pass-through and timing to simulate interest income and interest expenses at bank level for each macro scenario (baseline, geopolitical, recession). - Additional assumptions: - For fixed income securities assets: 100 percent pass-through from each scenario’s average of short-term and long-term interest rates. - For borrowings of banks: 50 percent pass-through from each scenario’s average of short-term and long-term interest rates, except for repo funding where 50 percent pass-through from each scenario’s short-term interest rate is assumed. - Accrual vs non-accrual exposures: - Non-accrual exposures do not earn interest income; net interest income can decline due to accumulation of non-accrual exposures. - Data sources used: - Bank level annualized NII from FINREP template F 16.01 (cumulating quarterly NII from 2024: Q3 to 2025: Q3). - Bank level exposures and deposits by segment and country from FINREP templates F 20.04 and F 20.06, and COREP as of 2025: Q3. - Repricing ladder from supervisory IRRBB template J 05.00. - Macro data from WEO and Haver Analytics. ### Interest rate pass-through econometric models (Box 6) - For Austria: - Dynamic panel regressions at bank-counterpart-instrument level based on OeNB’s MFI Interest Rate Statistics quarterly from 2016: Q1 onwards. - Liability-side regressions: sight retail deposits, term retail deposits, sight wholesale deposits, term wholesale deposits. - Asset-side regressions: consumer loans, mortgages, loans to SMEs, loans to large NFCs; each split by fixed rate and floating rate instruments. - Specification (first differences): - Δi_t,i_j = α_j + λ_j,c Δi_t−1,i_j + β0,j^ST Δi_t,c^ST + β1,j^ST Δi_t−1,c^ST + β1,j^LT Δi_t,c^LT + β1^LT Δi_t−1,c^LT + γ0,j GDPgrowth_t,c + γ1,j GDPgrowth_t−1,c + f_i + ε_i,j,t - Where Δ denotes quarterly changes; i_t,i_j is new business rate for bank i and segment j; ST is short-term rate (EURIBOR for EA, one-month interbank for non-EA); LT is long-term sovereign yield; GDPgrowth is year-on-year real GDP growth; f_i bank fixed effects. - For Germany and CESEE: - Time series aggregate regressions at quarterly frequency starting 2008 from ECB MFI interest rate statistics. - Segments: sight retail deposits, term retail deposits, sight wholesale deposits, term wholesale deposits, mortgages, loans to NFCs. - Specification in first differences analogous to Austria’s panel specification. _Italic line: Box 2. Austria: Probabilities of Default Models for Exposures to NFCs in Austria and Germany (continued), as provided in the source content._ ### 37. Shocks to interest rates simulated from the econometric models tend to result in a ### 37. Shocks to interest rates simulated from the econometric models tend to result in a ### Impact on Net Interest Income (NII) - In the recession scenario, shocks to interest rates tend to result in a compression of NII in the first year of the scenario. - In the geopolitical scenario, shocks to interest rates tend to result in an expansion of NII in the first year of the scenario followed by a compression of NII. - The pattern is largely driven by stronger pass-through to loan rates (especially floating rates) than to deposit rates, and downward stickiness of sight deposit rates. - This pattern is observed for exposures to Austria, Germany, and the CESEE countries. ### Shocks to deposit rates in Austria (Table 5) - Baseline scenario: interest rates on sight and term deposits for retail and wholesale remain almost constant throughout the entire scenario. - Geopolitical scenario: interest rates, especially for term deposits, increase in 2026 in line with the scenario’s increase in short-term and long-term rates, then decline particularly in 2028 following the decrease in market rates; by end-2028 deposit rates are about back to their starting point. - Recession scenario: an overlay is imposed to capture observed stickiness of sight deposit rates and to ensure sight deposit rates predicted do not turn negative and do not decline by more than the country aggregate of 0.3 percent for 2025:Q3. - As a result, cumulative changes over 3 years (all assumed to take place in the first year of the scenario as predicted by the econometric model): - Term deposits: cumulative decline of 1.2 percent for retail; 1.5 percent for wholesale. - Sight deposits: decline by a mere 0.3 percent. - Reported annual changes (in percentage points) from Table 5: - Baseline Scenario - Sight Deposits Retail: 2026 -0.01; 2027 0.00; 2028 0.00 - Term Deposits Retail: 2026 0.00; 2027 0.02; 2028 0.01 - Sight Deposits Wholesale: 2026 -0.1; 2027 -0.01; 2028 0.00 - Term Deposits Wholesale: 2026 -0.04; 2027 -0.01; 2028 0.00 - Geopolitical Scenario - Sight Deposits Retail: 2026 0.3; 2027 0.0; 2028 -0.3 - Term Deposits Retail: 2026 1.0; 2027 -0.2; 2028 -0.9 - Sight Deposits Wholesale: 2026 0.3; 2027 -0.2; 2028 -0.2 - Term Deposits Wholesale: 2026 1.1; 2027 -0.2; 2028 -1.0 - Recession Scenario - Sight Deposits Retail: 2026 -0.3; 2027 0.0; 2028 0.0 - Term Deposits Retail: 2026 -0.6; 2027 -0.6; 2028 -0.1 - Sight Deposits Wholesale: 2026 -0.3; 2027 0.0; 2028 0.0 - Term Deposits Wholesale: 2026 -0.8; 2027 -0.7; 2028 -0.1 - Given downward stickiness and lower pass-through to deposit rates, the recession scenario results in a significant compression of NII. ### Shocks to lending rates in Austria (Table 6) - Baseline scenario: interest rates on loans remain stable or increase very moderately, especially for new fixed rate loans. - Geopolitical scenario: interest rates rise sharply in 2026, by the same order of magnitude for the entire stock of floating rate loans and for new fixed rate loans, to a larger extent than the increase in deposit rates — explaining the increase in NII in 2026. In subsequent years, interest rates decline; cumulatively interest rates on loans are slightly below their starting point by end-2028 while deposit rates are about the same as their starting point by end-2028, implying NII is slightly below starting point by end-2028. - Recession scenario: interest rates on all floating rate loans, and on new originations of fixed rate loans decline steadily during each of the 3 years of the scenario. - Reported annual changes (in percentage points) from Table 6 by product type and rate type: - Consumer Loans / Mortgages (Floating Rate) - Baseline: 2026 0.0; 2027 0.1; 2028 0.0 - Geopolitical: 2026 0.2; 2027 0.2; 2028 0.1 - Recession: 2026 -0.7; 2027 -0.9; 2028 -0.2 - Consumer Loans / Mortgages (Fixed Rate) - Baseline: 2026 0.0; 2027 0.1; 2028 0.0 - Geopolitical: 2026 1.5; 2027 -0.3; 2028 -1.2 - Recession: 2026 -1.0; 2027 -0.9; 2028 -0.2 - SME Loans / Large Loans (Floating Rate) - Baseline: 2026 0.0; 2027 0.1; 2028 0.0 - Geopolitical: 2026 0.1; 2027 0.1; 2028 0.1 - Recession: 2026 -1.1; 2027 -0.9; 2028 -0.1 - SME Loans / Large Loans (Fixed Rate) - Baseline: 2026 0.0; 2027 0.0; 2028 0.0 - Geopolitical: 2026 1.2; 2027 -0.2; 2028 -1.2 - Recession: 2026 -1.0; 2027 -0.9; 2028 -0.1 ### Application to repricing ladders and NII simulation - Shocks to interest rates summarized in Tables 5-6 and Figures 33 a-f are applied to repricing ladders obtained from IRRBB J-templates for each bank. - Portfolio evolution is simulated under the static balance sheet assumption: portfolio compositions for interest-rate bearing assets and liabilities remain constant over the three years. - Shocks to deposit and lending rates by asset class and country of exposure are aggregated at the bank level based on exposures and applied to repricing ladder time buckets: - Buckets that reprice during that year receive the shock. - Buckets that have not repriced continue to receive the same interest income (assets) or bear the same interest expense (liabilities) for fixed rate instruments. - Floating rate loans all reprice at impact to the new interest rate. - For each year, interest income and expenses are aggregated at the bank level across repricing buckets. ### Market risk modelling - Market risk assessed for Austrian SIs and UniCredit’s subsidiary in Austria using a partial revaluation approach against the two market stress scenarios of the 2025 Euro Area FSAP. - Instruments at fair value (FVOCI and FVPL) revalued using bank-specific sensitivities reported in the ECB’s Short Term Exercise (STE): delta, gamma, vega. - Risk factors covered: commodity risk (CM), credit spread risk (CR), equity risk (EQ), interest rate risk (IR), and FX risk. FX shock corresponds to first-year FX depreciation in the adverse macro scenarios. - No market shocks applied in the baseline scenario. - Revaluation formula used: - ∆V = Σ_{j∈{CM,CR,EQ,IR,FX}} [(Delta_j^BB + Delta_j^TB) ∙ ∆ε_j + 0.5 ∙ Gamma_j ∙ (∆ε_j)^2 + Vega_j ∙ ∆σ_j] - Definitions: Delta_j^BB and Delta_j^TB are delta sensitivities for the banking book and trading book; ∆ε_j and ∆σ_j are shocks to the risk factor j and its volatility respectively. - A floor was applied to market risk losses: the maximum between i. the bank’s own fund requirements based on the Fundamental Review of the Trading Book (COREP C 91.00); and ii. 8 percent of the bank’s RWAs for market risk. - Note on banking book deltas: STE banking book deltas may include instruments categorized at amortized cost (AC) if the bank regularly calculates a fair value, potentially overestimating banking book market losses. ### Other P&L elements - Net Fees and Commission Income (NFCI): - Modelled with simple assumptions: NFCI assumed constant for each bank equal to the last end-of-financial-year data point (2024) as a share of assets for each year of the baseline. - In each adverse scenario, a haircut of 20 percent is applied to the ratio of NFCI to assets of the baseline if baseline NFCI was positive. If baseline NFCI was negative, the same NFCI applies to adverse scenarios. - Net Trading Income (NTI): - NTI projected alongside market risk revaluation. NTI equals 2024 annual value relative to total assets; in adverse scenarios a 25 percent haircut applied to these revenues for 2024 and kept at 25 percent for the 3 years of each adverse scenario. - All other P&L items set at their annual 2024 value relative to total assets and adjusted according to scenario inflation paths. - Income tax rate set at 30 percent. - Dividends: - For IFRS-reporting banks, the FSAP relied on OeNB’s dividend model based on bank-level regressions linking dividend payout ratios to bank characteristics and macro variables. - For other (small) banks, dividend-payout ratio assumed to be zero. ### Core solvency stress test results (aggregate and drivers) - Aggregate capital ratio remains above the minimum capital requirement in both adverse scenarios. - Baseline scenario: banking system remains well capitalized. - Decline in system-level CET1 ratio: - Recession scenario: decline of 465 basis points. - Geopolitical scenario: decline of 391 basis points. - Contributions to decline (relative to baseline) and ranking of drivers: - Credit risk: top driver — 420 basis points in the geopolitical scenario; 404 basis points in the recession scenario. - Decline in NII: second largest contributor — 171 basis points in the geopolitical scenario; 329 basis points in the recession scenario. - Market risk of SIs: contributes 138 basis points in the geopolitical scenario; 146 basis points in the recession scenario. - Timing: contributions to the capital ratio decline occur mostly upfront in 2026 due to the expected loss approach for credit risk, causing sharp increases in provision outflows and RWA from accumulation of defaulted exposures. - NII dynamics: - Recession scenario: strong NII decline from 2026 onwards due to compression of interest rate spreads. - Geopolitical scenario: NII initially increases in 2026 then declines in 2027 and 2028. - Market shock and NFCI haircut are assumed to occur in 2026 and carried over in 2027 and 2028. ### SIs versus LSIs and breach statistics - Under the two adverse scenarios, SIs experience larger declines in capital ratios than LSIs. - Decline for SIs: 424 basis points in the geopolitical scenario; 500 basis points in the recession scenario. - Decline for LSIs: 326 basis points in the geopolitical scenario; 396 basis points in the recession scenario. - Reasons: - SIs experience larger credit risk and larger decline in NII (including from foreign exposures) and also experience the one-off global market shocks. - SIs pay less taxes and distribute less dividends in adverse scenarios relative to the baseline, which mitigates some decline in capital ratios; this effect does not operate for LSIs. - Breaches: - Only a few small LSIs breach their Total SREP Capital Requirement (TSCR). - Twelve banks dip into their additional buffers (CCB, OSII and systemic risk buffers), among which three are SIs. - Three (one) small LSIs accounting for 0.36 percent (less than 0.2 percent) of banking assets breach their TSCR in the geopolitical scenario (in the recession scenario). - Twelve banks (among which three SIs) accounting for 9.5 percent (9.1 percent) of total banking assets of the sample breach their requirements including buffers in the geopolitical scenario (in the recession scenario). - Weighted average gaps to the requirement for those banks remain small at 1.4 and 2.2 percent of risk weighted assets on average, respectively in the geopolitical scenario and the recession scenario. *Source: IMF staff calculations and OeNB.* ### 52. This section describes the analysis quantifying second-round effects of the solvency ### 1autea2026007 - 52. ### Second-round effects of solvency stress tests via IPS recapitalizations - Mutual support mechanism: IPS ex-ante funds and ex-post recapitalizations by banks with a “capital surplus” can be relied upon to recapitalize member banks that have become undercapitalized, up to their TSCR. - Contribution to IPS: Banks that belong to the IPS have contributed 0.5 percent of their total RWA to the IPS to pre-fund the IPS. - Ex-post contributions: Member banks can be asked to contribute ex-post to recapitalization as long as their capital ratios remain above their own TSCR plus a buffer of 10 percent. - Two hurdle rates analyzed: - Hurdle H1: TSCR (actual hurdle rate of the IPS of the decentralized sector for recapitalization by IPS funds). - Hurdle H2: TSCR plus buffers (CCoB, O-SII and systemic risk buffers) — hypothetical illustrative exercise; not applied in practice in Austria for IPS activation. ### Findings on recapitalization sufficiency - Under hurdle H1 (TSCR): - Each individual regional IPS ex-ante fund is sufficient to recapitalize all banks of the decentralized sectors to their TSCR. - In each adverse scenario, IPS ex-ante funds of all individual regions remain in surplus after recapitalizations. - After recapitalization, only a small portion of the system-wide ex-ante IPS fund is relied upon. - Under hurdle H2 (TSCR plus buffers) — hypothetical: - Several regions experience deficits of ex-ante funds and must resort to ex-post recapitalizations. - Three regions out of eight lack sufficient ex-ante funds to recapitalize their member banks. - Surplus banks in each region must contribute ex-post to recapitalizing banks in deficit. - Each of the eight regions has sufficient capital among its members to achieve complete recapitalizations with ex-post contributions, without resorting to contributions from members outside of the region. ### Quantitative indicators (undercapitalized banks, hurdle definitions) - Hurdle Rate = Minimum Total SREP Capital Requirement (TSCR). - Hurdle Rate = Minimum buffer requirement (TSCR + buffers). - Under categories shown for CET1 (two scenarios): - CET1 - Geopolitical Scenario: # banks below requirement = 3; Market share of undercapitalized banks = 0.36%; # banks below buffers = 12; Market share of banks below buffers = 9.5%; Weighted average capital ratio gap = 1.4%. - CET1 - Recession Scenario: # banks below requirement = 1; Market share of undercapitalized banks = less than 0.2%; # banks below buffers = 12; Market share of banks below buffers = 9.1%; Weighted average capital ratio gap = 2.2%. ### Sensitivity analysis and concentration risks - Large exposures: - Top ten largest exposures to individual clients or groups reach more than six times the CET1 of Austrian banks at the aggregate before CRM. - After netting out credit risk mitigation measures (CRM), the top ten largest exposures reach only 10 percent of aggregate CET1. - CRM definition: collateralization, guarantees and insurance, netting, covenants and derivatives. ### Exposures to NBFIs - Aggregate large banks’ exposures to NBFIs: - Before CRM: about 100 percent of their CET1 (either through direct exposures or indirect exposures that are part of other banking groups). - After CRM: exposures are 30 percent of aggregate CET1 of Austrian banks. - COREP reporting insights: - Aggregated 8 banks (share of CET1) large exposures before/after exemptions and CRM: - Top 5: (1) 3.4; (2) 0.1 - Top 10: (1) 6.3; (2) 0.1 - Top 20: (1) 11.7; (2) 0.2 - Aggregated 6 SIs+UniCredit Austria (share of CET1) before/after exemptions and CRM: - Group level NBFI (1): 0.6; (2): 0.1 - Individual NBFI (1): 0.3; (2): 0.2 - Total (1): 1.0; (2): 0.3 - Large exposures to NBFIs by NACE: - Group level NBFI (3): 0.50; (4): 0.15 - Individual NBFI (3): 0.23; (4): 0.02 - Total (3): 0.73; (4): 0.18 - NACE composition: about 73 percent of these NBFIs belong to NACE sector K64 and 17 percent to K66. - OeNB balance sheet item (BSI) data: exposures are primarily to OFIs, with a substantial portion being part of a banking group, particularly in domestic loans and domestic equity exposures. ### Sovereign-bank nexus - Exposures to general government: - SIs: 15.7 percent of assets. - LSIs: 7.8 percent of assets. - All banks: 13.2 percent of assets. - Geographical composition of government exposures: - SIs: 21.4 percent to the Austrian sovereign; larger share to CESEE sovereigns. - LSIs: about 69.6 percent to the domestic sovereign. - Deposit funding costs: - Deposit costs in Austria are mostly sensitive to short-term rates and almost not affected by long-term rates. - No statistically and economically significant association between interest rates on sight deposits and long-term rates. - Term deposit rate has a short-term pass-through of about 10 percent to long-term rates. - Similar conclusions broadly hold for deposit costs in Germany and CESEE countries. - Analysis did not assess sensitivity of borrowing costs of banks to government bond yields. ### Stress tests of non-financial private sectors (overview) - Stress tests performed under baseline and two adverse scenarios (geopolitical and recession), consistent with bank solvency stress tests. - Methodologies: - Large NFCs: corporate stress test model (Box 6), results aggregated using indebtedness weights. - Households: household micro-macro model (Box 4), used to simulate mortgage and consumer loan PDs and for counterfactual policy analysis. ### Solvency and liquidity stress tests of publicly listed NFCs - Debt-at-risk (ICR < 1): - Baseline: increases from 15 percent of total debt to 21 percent (after scenarios). - In two adverse scenarios: increases very substantially to more than 60 percent of total debt after two years and declines thereafter. - PD elasticities (aggregate): - Geopolitical scenario: about 1.69 in 2026 and 1.77 in 2027. - Recession scenario: about 1.47 in 2026 and 1.7 in 2027. - These elasticities are above average for non-real estate sectors from the credit register econometric model, but below those for real estate related sectors. - Liquidity needs: - Baseline: share of debt of firms with liquidity needs declines substantially from initial value. - Adverse scenarios: share remains high in 2026 and 2027 at 35-40 percent, implying increased liquidity needs and further leveraging. ### Household vulnerabilities and counterfactual analysis of borrower-based instruments - Comparative household indebtedness: - Debt-to-income ratios of Austrian households are lower than in Germany, France, Italy and Spain for the bottom 50 percent and for the top 25th percentile. - Debt-to-financial assets ratios: very low in Austria across income groups; similar to Germany; significantly lower than France, Italy and Spain. - Counterfactual: binding LTV limit at origination of 90 percent on newly originated housing loans (aligned with KIM-V regulation, 2022). - Implementation assumption: exclude from survey population households with LTV at origination above 90 percent; exposures rescaled to maintain constant total balance sheet of banks. - Effect on annual PDs (Table 7; annual default rates, benchmarked to EBA risk dashboard 2025:Q3): - No LTV limits (Annual default rates): - Baseline: 2025 1.27%; 2026 1.26%; 2027 1.25%; 2028 1.24%; 2029 1.23%; 2030 1.23% - Geopolitical scenario: 2025 1.27%; 2026 1.35%; 2027 1.79%; 2028 2.12%; 2029 1.98%; 2030 1.61% - Recession scenario: 2025 1.27%; 2026 1.42%; 2027 2.18%; 2028 2.67%; 2029 2.40%; 2030 1.81% - Only LTV at origination < 90 percent (Annual default rates): - Baseline: 2025 0.73%; 2026 0.73%; 2027 0.72%; 2028 0.71%; 2029 0.71%; 2030 0.71% - Geopolitical scenario: 2025 0.73%; 2026 0.78%; 2027 1.01%; 2028 1.22%; 2029 1.15%; 2030 0.93% - Recession scenario: 2025 0.73%; 2026 0.82%; 2027 1.22%; 2028 1.54%; 2029 1.39%; 2030 1.04% - Impact on PDs: With no LTV limits, annual PDs peak in 2028 with an increase of 0.85 percentage points (geopolitical) and 1.4 percentage points (recession). With LTV limit, increases are more contained to 0.5 percentage points (geopolitical) and 0.82 percentage points (recession). - Cumulative impact on defaulted exposures (5-year extrapolation, assuming no cures and no write-offs; Table 8): - Cumulative difference in share of defaulted exposures after 5 years between no-LTV and LTV<90%: - Baseline: about 3 percentage points, corresponding to €5.6 billion reduction in stock of defaulted exposures. - Geopolitical scenario: about 4-4.6 percent, corresponding to €7.5 billion (geopolitical) and €8.7 billion (geopolitical stated twice in source) of reduction in stock of defaulted exposures after five years when LTV limits of 90 percent are in place at origination. _https://www.imf.org/-/media/files/publications/cr/2026/english/1autea2026007.pdf_ ### 64. Liquidity risks are assessed through structural analysis on the Liquidity Coverage Ratio ### 64. Liquidity risks are assessed through structural analysis on the Liquidity Coverage Ratio (LCR) ### Metrics and scope - Liquidity risks are assessed through structural analysis on the Liquidity Coverage Ratio (LCR), Net Stable Funding Ratio (NSFR), asset encumbrance, and funding concentration. - LCR measures banks’ ability to withstand short-term outflows by drawing on high-quality liquid assets (HQLA). - NSFR gauges structural financing vulnerabilities by matching available stable funding against required stable funding for long-term, illiquid assets. - Asset encumbrance analysis evaluates the proportion of collateralized assets to reveal structural constraints on banks’ capacity to mobilize unencumbered collateral during stress events. - Funding concentration metrics measure diversification of funding and exposures to concentrated fund outflows. ### Aggregate projection examples (difference in share of cumulative defaulted exposures; no LTV limits minus 90 percent LTV limit) - Baseline scenario: - 2025: 0.54% - 2026: 1.06% - 2027: 1.57% - 2028: 2.07% - 2029: 2.55% - 2030: 3.02% - Geopolitical scenario: - 2025: 0.54% - 2026: 1.10% - 2027: 1.85% - 2028: 2.69% - 2029: 3.44% - 2030: 4.03% - Recession scenario: - 2025: 0.54% - 2026: 1.14% - 2027: 2.05% - 2028: 3.09% - 2029: 3.99% - 2030: 4.64% ### Difference in cumulative defaulted exposures (in percent) (no LTV limits minus 90 percent LTV limit) - Baseline scenario: - 2025: 1.02 - 2026: 1.99 - 2027: 2.95 - 2028: 3.87 - 2029: 4.77 - 2030: 5.65 - Geopolitical scenario: - 2025: 1.02 - 2026: 2.07 - 2027: 3.47 - 2028: 5.03 - 2029: 6.43 - 2030: 7.54 - Recession scenario: - 2025: 1.02 - 2026: 2.13 - 2027: 3.84 - 2028: 5.79 - 2029: 7.46 - 2030: 8.68 - Source for these estimates: HFCS 2021 and IMF staff estimates ### LCR and cash-flow stress testing approach - LCR stress tests evaluate banks' capacity against 30-day liquidity drains using two LCR scenarios parameterized with larger asset haircuts and higher secured and unsecured outflows than the regulatory baseline. - Scenarios are based on the 2025 Euro Area FSAP LCR scenarios, re-calibrated to capture structural characteristics of the Austrian banking system. - Cash-flow stress tests examine banks’ liquidity positions over longer periods up to 12 months to identify resilience and vulnerability over extended horizons; scenarios are based on the 2025 Euro Area FSAP with key changes on deposit outflow assumptions to reflect Austrian funding dynamics. - Second-round effects from the Institutional Protection Scheme (IPS) and liquidity associations of a major decentralized sector are modelled to capture mutual contractual obligations, liquidity reallocation, and pre-contributed IPS funds to quantify aggregate system capacity to absorb liquidity shocks. ### Structural liquidity risks — key findings LCR - Under the standard regulatory scenario, the weighted average LCR of Austrian banks is 185 percent as of September 2025. - EU/EEA benchmark: 163 percent in December 2024. - SIs average LCR: 162 percent. - LSIs average LCR: 240 percent. - 34 LSIs show LCR below 100 percent; these participate in regional liquidity associations subject to liquidity waivers. Consolidation at association level results in compliance with 100 percent requirement. HQLA composition - Cash and central bank reserves constitute 39 percent of HQLA for both SIs and LSIs. - Government and public sector debt: 39 percent of SI’s HQLA; 29 percent of LSI’s HQLA. - LSIs allocate a greater proportion to other Level 1 assets (e.g., extreme high-quality covered bonds). - Exposures to Level 2 assets remain small across the sector. - A mandatory liquidity upstreaming arrangement exists within the major decentralized sector, transferring part of small regional banks’ L1 assets to central regional institutions. Deposit and outflow profiles - Retail deposits represent 57 percent of LSI liabilities and 47 percent of SI liabilities (unweighted terms). - "Other products and services" outflows: 16 percent for SIs’ total unweighted outflows; 13 percent for LSIs. - Austrian banks assign an average outflow rate of 5.5 percent to these exposures for LCR calculations under the regulatory scenario. Currency-specific liquidity buffers - Only EUR and CZK account for more than 5 percent of total outflows. - Weighted average LCR by currency: - EUR: 183 percent - CZK: 157 percent NSFR - Weighted average NSFR: 140 percent for SIs and 132 percent for LSIs. - Basel III requirement / EU/EEA benchmark: 127 percent. - Available Stable Funding (ASF) composition: retail deposits account for 50 percent of SI funding and 51 percent of LSI funding. - Liabilities from undetermined counterparties: 16 percent for SI and 15 percent for LSI. - Capital: 13 percent for SI and 15 percent for LSI. - LSIs hold a greater share of funding maturing within six months than SIs, exposing LSIs to higher rollover risk. - Loan portfolios: principal component of Required Stable Funding (RSF) for SIs and LSIs. Asset encumbrance - Asset encumbrance ratio (share of total assets pledged as collateral): - SIs: 14 percent - LSIs: 24 percent - EU/EEA average: 24.1 percent. - LSIs report an encumbrance rate of 34 percent on loan portfolios; SIs report 18 percent on loan portfolios. - Austrian banks encumber loans more frequently than debt securities, contrasting with broader European practice. Funding concentration - Funding concentration ratio (share of funding from 10 largest providers to total funding) — median: - SIs: 6.3 percent - LSIs: 8.8 percent - Euro Area FSAP median for lenders and universal banks: 10 percent. - 75th percentiles: SI 9 percent; LSI 17 percent. - 90th percentiles: SI 13 percent; LSI 31 percent. ### LCR stress tests — assumptions and recalibrations - Two adverse scenarios built on 2025 Euro Area FSAP scenarios; assume sovereign stress triggers larger valuation haircuts on HQLA and elevated secured and unsecured outflows. - Re-calibrations for Austrian context: - Retail deposits subject to higher outflows: COREP sub-categories average outflow rates under regulatory reporting are 14.9 percent and 19.9 percent; LCR scenarios adopt 14.9 percent and 19.9 percent (instead of regulatory lower limits 10 percent and 15 percent). - "Other products and services" unsecured outflows: regulatory scenario average reported by banks ≈ 5.5 percent; EA FSAP assumed 100 percent. Historical wholesale deposit monthly outflows over 20 years identify 75th percentile ≈ 25 percent and 90th percentile ≈ 40 percent; these two figures are used for calibration of this category under the two LCR scenarios. ### LCR stress tests — results - Weighted average LCR under adverse scenarios: - Scenario 1: 149 percent (weighted average) - SIs: 126 percent (average) - LSIs: 203 percent (average) - Scenario 2: 133 percent (weighted average) - SIs: 109 percent (average) - LSIs: 190 percent (average) - Currency-specific weighted average LCR under scenarios: - Scenario 1: - EUR: 151 percent - CZK: 132 percent - Scenario 2: - EUR: 137 percent - CZK: 121 percent - Vulnerabilities: - Scenario 1: banks with LCR below 100 percent are only LSIs, representing 7 percent of total banking sector assets (including banks with 2 percent market share operating within regional liquidity associations with waivers). - Scenario 2: several SIs also fall below 100 percent; total asset share of banks below 100 percent rises to 35 percent. - Historical stress testing (SIs, October 2020 to September 2025 monthly COREP): - Weighted average LCR stays consistently above 100 percent under regulatory and two adverse scenarios. - Standard deviations: 15 percent under Scenario 1 and 13 percent under Scenario 2. - Note: slight downward trend in reported LCR over sample period. ### Cash-flow based liquidity stress tests — design and results - Scenarios: - General risk aversion scenario: aggregate shocks to macroeconomic confidence, valuation losses in counterbalancing capacity (CBC), limited access to funding markets, systemic liquidity facility drawdowns. - Idiosyncratic bank run scenario: confidence shocks for individual banks leading to severe and sustained funding outflows; simulates credit losses or unprofitable business models. - Analysis focuses on net liquidity position = cumulated CBC minus cumulated net outflows across maturity buckets from overnight to 12 months. - Deposit outflow assumptions aligned with LCR tests for 30-day cumulative outflow rates; beyond 30 days cumulative outflows anchored to original EA FSAP assumptions on 12-month horizon. Cash-flow stress test results - General risk aversion scenario: - Aggregate net liquidity position remains positive across all time buckets. - Total market share of banks experiencing shortfalls is 10 percent at 12 months; these are LSIs, about half (market share) operating within regional liquidity associations with waivers. - Idiosyncratic scenario: - System maintains positive aggregate position up to three months before turning negative. - By 12 months, market share of banks with shortfalls reaches 81 percent, driven primarily by several large SIs facing liquidity gaps after three months. - Currency-specific outcomes: - Under general risk aversion, both EUR and CZK net liquidity positions remain positive. - Under idiosyncratic stress: - EUR net position moderates toward zero by six months. - CZK net position becomes negative after three months; by the end of 12 months the cumulated negative net liquidity position equals 27 percent of total banking sector assets in absolute terms. ### Second-round effects: IPS and liquidity associations - Modeling incorporates mutual contractual obligations under the Institutional Protection Scheme (IPS) and liquidity associations of the major decentralized sector to capture network dynamics. - Member banks facing liquidity shortages can receive funding via liquidity reallocation within associations and by using pre-contributed IPS funds. - This holistic perspective identifies channels of systemic risk transmission and quantifies the aggregate capacity of the system to absorb liquidity shocks. *Source: HFCS 2021 and IMF staff estimates* ### 83. Further liquidity stress tests incorporating second-round effects from the IPS and ### Further liquidity stress tests incorporating second-round effects from the IPS and liquidity associations ### Overview and purpose - Standard liquidity stress testing methodologies evaluate banks independently and overlook mutual liquidity support mechanisms present in the Austrian banking system. - Many banks operate under shared contractual IPS arrangements within liquidity associations, obliging mutual liquidity support in stress periods. - Analysis models these mechanisms within a major decentralized sector comprising 3 SIs and 274 LSIs, accounting for 39.2 percent market share in the Austrian banking system as of September 2025. - The IPS features a two-layer structure consisting of eight regional liquidity associations and one federal liquidity association; regional associations contain small retail banks and a central regional institution, while all central regional institutions and a central national institution form the federal association. ### Mechanics: four-stage mutual support modeled in LCR and cash-flow stress tests - Two channels of support in stress periods: - Receiving excess liquidity from surplus banks. - Drawing down IPS funds built through prior member contributions. - Four sequential stages (applied in both LCR and cash-flow tests where relevant): 1. Intra-group liquidity reallocation: surplus banks transfer excess liquidity to deficit banks within each local liquidity association. 2. Cross-group liquidity reallocation: sharing excess liquidity across all liquidity associations if deficits remain. 3. Drawdown of pre-contributed IPS funds within each local liquidity association. 4. Pooling residual IPS funds across associations to cover remaining deficits. - Application in tests: - LCR stress test: sequential support provided to member banks with LCR below the 100 percent threshold. - Cash-flow stress test: support provided to member banks facing a negative net liquidity position within each time bucket. - Modeling assumptions and notes: - Each member bank’s contribution is approximately equal to 0.5 percent of its risk-weighted asset. - Excess liquidity (LCR test) = liquid assets held above amount required for 100 percent regulatory threshold. - Excess liquidity (cash-flow test) = positive cumulative net liquidity position within a specific time bucket. - When multiple surplus banks exist, model allocates contributions in proportion to each bank's available excess; when multiple deficit banks exist, model assumes banks with larger market shares receive support first. - Ex-post discretionary contributions by the IPS risk council are not modeled; focus is on deterministic channels: reallocation of excess liquidity and drawdown of pre-contributed "ex-ante" IPS funds. - Model assumes instantaneous transfers and free reallocation of liquid assets; acknowledges these simplifications (see below for limitations). ### LCR stress test findings - Scenario definitions and aggregate excess liquidity: - Scenario 1 (milder): total excess liquidity = 7.0 percent of the sector’s total RWA. - Scenario 2 (severe): total excess liquidity = 4.2 percent of the sector’s total RWA. - Effectiveness by stage: - Scenario 1: - Stage 1 (intra-group reallocation) is sufficient to restore all member banks falling below the 100 percent regulatory threshold. - Stage 1 utilizes 13.4 percent of the total excess liquidity. - Scenario 2: - Stage 1 and Stage 2 (cross-group reallocation) restore all member banks’ LCR by consuming 74.6 percent of the total excess liquidity. - Two liquidity associations exhaust internal buffers and require Stage 2 support; rescuing only these two associations consumes 29.1 percent of the total excess liquidity, indicating concentration of risks under severe stress. - Interpretation: - Mutual support mechanisms can restore regulatory compliance across member banks under both scenarios, but severe shocks reveal capacity divergence across associations and concentrated vulnerabilities. ### Cash-flow stress test findings and time dynamics - General patterns: - Support resolves member bank shortfalls in milder General Risk Aversion (GRA) scenario across all time buckets. - Under the idiosyncratic scenario, Stage 1 and Stage 2 fully offset shortfalls in the first three months, but continuous buffer erosion produces a steep cliff effect in the 4–5-month bucket. - Specific outcomes: - GRA scenario: - Stage 1 absorbs all liquidity gaps for the initial two months; Stage 2 activates starting the third month. - Idiosyncratic scenario: - Stage 1 and Stage 2 offset shortfalls up to month 3; by the end of month 4 banks representing over 73 percent of total sector assets experience liquidity shortfalls; by the end of month 12 this reaches over 96 percent. - Contribution ranking and contributor impact: - From recipients’ perspective, total liquidity received maintains ranking: Stage 1 > Stage 2 > Stage 3 > Stage 4 across scenarios and horizons. - From contributors’ perspective, initial buffers are primarily consumed by stress outflows; reallocating liquidity to deficit banks is a marginal drain—mutual support does not appear to significantly burden healthy surplus banks relative to the shock’s direct effects. ### Reverse stress tests and inter-temporal risk transmission - Reverse stress test design: - Continuum of 50 scenarios ranging from a baseline (Scenario 1 for LCR; GRA for cash flows) to an extreme where haircuts and runoff rates double. - Findings: - Reverse LCR stress tests: IPS and liquidity association support consistently reduces the market share of banks with LCR below the regulatory threshold across all severity levels. - Reverse cash-flow stress tests: support reduces liquidity gaps across all severities for short-term horizons, and for small to moderate severities over long-term horizons. - Inter-temporal transmission risk: - At longer horizons (six months or longer) and under high shock severity, the market share of banks facing shortfalls can be larger when support is provided than when it is not. This occurs because early draws on surplus banks deplete buffers, leaving them more exposed later if central bank intervention is absent. - This creates an inter-temporal risk transmission channel: exhausting system capacity early may limit ability to absorb subsequent shocks. ### Limitations and potential vulnerabilities of the modeled framework - Simplifying assumptions that may mask risks: - Instantaneous liquidity transfers without decision lags; in practice, administrative frictions can delay support and amplify systemic distress via missed payment obligations. - Assumption that liquid assets can be reallocated without costs. In practice: - Stages 3 and 4 (pre-contributed IPS funds) are cash and transferable with minimal costs. - Stages 1 and 2 rely largely on HQLA assets; monetizing these via forced liquidation or collateral pledging may trigger further asset haircuts and systemic liquidity spirals. - Implication: model results show potential effectiveness but should be interpreted considering operational frictions and market-run dynamics. ### Liquidity to solvency interactions: forced liquidation sensitivity analysis - Methodology: - Builds on Chapter 2 of October 2023 GFSR; assumes banks sell government bonds held at amortized cost (AC) when other counterbalancing capacity components are insufficient to cover funding gaps at two-week and 30-day horizons. - Analysis for two cash flow scenarios: general risk aversion and idiosyncratic. - Valuation effects calibrated based on term premium shocks occurring in 2026: Q1 in the recession scenario; modified duration approach using remaining maturities from COREP. - Analysis performed solely for SIs; top eight sovereign exposures per bank in COREP considered (24 country exposures). - Term premium shocks (Table 9; in percentage points) include: - AT (Austria) 1.46 - FR (France) 2.28 - BE (Belgium) 2.57 - ES (Spain) 4.79 - DE (Germany) 1.42 - SK (Slovakia) 3.40 - IT (Italy) 4.58 - IE (Ireland) 2.43 - FI (Finland) 1.27 - NL (Netherlands) 1.31 - CZ (Czechia) 2.72 - RO (Romania) 2.72 - SK (Slovakia) 3.40 - HR (Croatia) 2.72 - HU (Hungary) 2.86 - US (United States) 2.29 - CA (Canada) 3.84 - RS (Serbia) 2.72 - BG (Bulgaria) 2.72 - PL (Poland) 4.94 - BA (Bosnia and Herzegovina) 2.72 - JP (Japan) 1.3 - NO (Norway) 3.64 - GB (United Kingdom) 2.35 - Key solvency results (Table 10 summary): - General Risk Aversion Scenario: - Average solvency impact from forced sell-off of AC government bonds = 1.0% of RWA over two weeks and 1.3% of RWA over 30 days. - In percent of CET1: average 5.6% over two weeks and 7.7% over 30 days. - Standard deviations: 2.9% and 3.7% of RWA; 16.3% and 21.2% of CET1 respectively. - Idiosyncratic bank run scenario: - Average solvency impact = 2.4% of RWA over two weeks and 3.9% of RWA over 30 days. - In percent of CET1: average 13.7% over two weeks and 22.2% over 30 days. - Standard deviations: 3.7% and 5.0% of RWA; 20.7% and 28.1% of CET1 respectively. - Memo: Aggregate cumulative net outflows in billions of euros: 23 (2 weeks) and 35 (30 days) for GRA; 75 (2 weeks) and 109 (30 days) for idiosyncratic scenario. - Interpretation: - Liquidity shortages resulting from bank runs would lead to significant additional depletion of bank capital, with materially larger impacts under the idiosyncratic bank run scenario. - Practically, central bank liquidity provision would reduce the need for fire sales and resulting solvency losses; however, the forced-sale exercise is informative for resilience assessment. ### Interconnectedness and contagion analysis - Interbank network characteristics: - Austrian interbank market displays complex interconnectedness centered around large SIs. - Interbank transactions within the decentralized sector are mostly between smaller and larger members. - Large SIs have substantial exposures to each other across decentralized and non-decentralized sectors. - Non-decentralized sector has large exposure to the decentralized sector. - Complex protection schemes exist to prevent contagion; focus of analysis is on protection schemes in the decentralized sector. - Contagion modeling approach: - Based on CoMAP framework calibrated to Austrian interbank market. - Two main contagion channels modeled: 1. Credit loss channel: defaults propagate credit losses via interbank exposures. 2. Funding/liquidity channel: losses create funding shortfalls that must be backed by HQLA or sold at fire-sale discounts. - Default/distress condition: impacted bank’s capital falls below regulatory threshold after credit losses, fire sales, or liquidity pool adjustments. - Simulation iterates until convergence (no new defaults). - Data: interbank bilateral exposures as of December 2024 used for calibration. *IMF staff analysis as presented in the source PDF.* ### 96. Results are reported from two different (systemwide contagion and individual ### B. Bank Contagion Analysis ### Model approach and metrics - Two angles reported: systemwide contagion (contagiousness measures how one bank’s default affects the whole system’s capital and liquidity position) and individual vulnerability (measures how (on average) a bank’s individual capital is affected by other banks’ hypothetical defaults). - Analysis focuses on potential cascade effects within Austria only; potential cross-border effects are not modelled. - Contagion amplification defined as losses induced by a bank’s simulated default in the initial round vis-à-vis those occurring in all successive rounds. - Vulnerability amplification defined as the ratio of the losses experienced by an individual bank across all simulations relative to losses in successive rounds. - Distress threshold defined as CET1 capital plus all buffers over the regulatory minimum. - Net liquidity pool (훾) = Liquidity Buffers minus Net Liquidity Outflows (equals Net Liquidity Position above the value required by an LCR ratio of 1). - Firesales haircuts (훿) calibrated to be 35 percent. - Discount rates (휚) calculated as the ratio of total interbank funding is short term (lower than 30 days maturity) for each bank and averaged around 0.3 in the sample. - Pool of assets available for fire sales (휃) = counterbalancing capacity minus non tradable assets eligible for central banks minus HQLA holdings. - Available HQLA assumed usable for central bank support. ### Scenario design and major assumptions - Three scenarios considered: - Baseline calibration when the IPS protection scheme was in place. - Counterfactual scenario when the IPS scheme was missing. - CESEE loss scenario where the Austrian banking system suffers losses from exposures from CESEE countries. - Bilateral exposures matrix used December 2024 data. - Loss Given Defaults (LGD, 휆) assumed equal to the ratio of Gross Exposure minus Collateralized Lending to Gross exposures for each individual exposure. - On average, 휆 was found to be between 0.9 to 1 for the non-decentralized banks (i.e., most interbank exposures are uncollateralized). - Simplifying assumption for IPS role in decentralized sector: if either lender or borrower was among the smaller LSI banks in the decentralized sector, set 휆 to 0.3 for those exposures (assumes IPS covers a large portion of potential losses for smaller members). If default impacted Tier 2 and Tier 3 banks in the decentralized sector, used the calculated bilateral 휆s (which lie between 0.9 and 1). - In the counterfactual scenario (without IPS) used the 휆s calculated for all banks (on average around 0.9) for the smaller IPS member LSIs too. - CESEE loss scenario: shocked the initial capital stock of seven SI banks with large CESEE exposures (including Russia) to nonfinancial corporations and households. The overall capital shock to the system amounted to 8.7 billion euros. - Major assumptions summarized in Table 11 (not reproduced here). ### Key quantitative summary statistics (as reported) - Average, Weighted Average, median, Standard error (Baseline (with IPS) and CESEE loss): 0.88; 0.95; 1.00; 0.27. - Average, Weighted Average, median, Standard error (Without IPS): 0.99; 0.99; 1.00; 0.08. - All scenarios: Average; Weighted Average; median; Standard error: 0.33; 0.25; 0.27; 0.29. - CESEE loss — Funding Shortfall (FS) / Loss Given Default (LGD): Shock to Capital in bn euros for 7 SI banks: 8.74. ### Main findings - Contagion risks are highly concentrated and originate from a few SIs (Table 12). - Vulnerabilities from interconnectedness are found to be limited in the Austrian banking system. - The system is resilient to hypothetical defaults; only a limited number of banks would create contagion risks. - The most contagious bank’s hypothetical default would induce a system-wide loss reaching the total capital of the whole system. This is around four times that of the loss resulting from the second most contagious bank and around ten times the average of the most contagious twenty banks’ impact. - Amplification through cascade effects is large in the case of the most contagious bank, while less important for other banks. - The most vulnerable banks were smaller in size. - Systemic risk map found no banks to be contagious and at the same time vulnerable (Figure 31). - Overwhelming majority of banks were found neither contagious nor vulnerable. - A few banks were found contagious, but their HQLA holdings, liquidity pools available for fire sales and capital prevent them from being vulnerable. - Most vulnerable LSIs with relatively tight liquidity or capital positions are small relative to the whole system. ### Role of the IPS scheme - IPS scheme in the decentralized sector generally dampens vulnerabilities and may amplify interconnections in case of stress to larger banks. - Counterfactual analysis highlights that the scheme generally lowers the vulnerability of participating institutions. - Smaller LSIs’ potential capital losses arising from stress of other IPS member banks are generally dampened by the scheme. - For larger banks, the IPS scheme may strengthen interconnections due to their significant size, making it more likely that stress in these banks will increase spillover effects to lower-tier banks. - This result is based on the assumption that some of the losses may be covered by fire sales (which are not valued at mark-to-market) and on the assumption that in practice, larger banks would have to contribute relatively more than their pro-rata share than smaller ones. - These losses are muted by the non-modelled regulatory limits in practice. *Source: IMF staff analysis as presented in the Austria bank contagion section of the provided chapter.* ### 101. Large exposures of the SIs in the CESEE region account for a significant share of ### Large exposures of the SIs in the CESEE region account for a significant share of ### Key findings - Our comprehensive data analysis of large exposures found that 18.5 percent of SIs total assets (before credit mitigation) are exposed to the CESEE region. - This exposure comes mostly from assets held in government and central bank liabilities and around 3 percent of total assets from non-financial corporations. - The largest country-level exposures (Grand Total, percentage of total assets) are: - Czech Republic: 7.2 - Hungary: 3.0 - Romania: 2.7 - Slovakia: 2.0 - Croatia: 1.5 - Poland: 0.9 - Serbia: 0.5 - Slovenia: 0.2 - Ukraine: 0.2 - Bosnia and Herzegovina: 0.1 - Albania: 0.1 - Bulgaria: 0.0 - Montenegro: 0.0 - North Macedonia: 0.0 - Grand Total (all CESEE): 18.5 - Data coverage note: figures are calculated for end of December 2024 and include 7 SI banks and UniCredit’s subsidiary in Austria. Although UniCredit is a subsidiary of an Italian bank, most of the group’s CESEE exposures are operated through the Austrian subsidiary. ### Stress-test exercise and network effects - An exercise where capital is lowered by potential losses on NFC exposures from CESEE demonstrates that these losses would be amplified in the network. - Amplification mechanism: most losses would affect the capital of the most contagious banks which then would cascade through the interbank market. - System vulnerability: vulnerability of the system is less affected because the directly impacted banks are generally not the most vulnerable ones. - Methodological note (footnote 59): Large exposures were then multiplied by five times the probabilities of default used by OeNB in its stress testing model (ARNIE). ### Table 13: Austria — SIs’ Large Exposures to CESEE Countries (percentage of total assets) - Albania — Grand Total: 0.1 - General governments: 0.1 - Bosnia and Herzegovina — Grand Total: 0.1 - Central banks: 0.1 - Bulgaria — Grand Total: 0.0 - Croatia — Grand Total: 1.5 - Central banks: 0.5 - General governments: 0.8 - Non-financial corporations: 0.2 - Czech Republic — Grand Total: 7.2 - Central banks: 3.1 - Credit institutions: 0.2 - Financial corporations other than credit institutions and investment firms: 0.2 - General governments: 3.0 - Non-financial corporations: 0.8 - Hungary — Grand Total: 3.0 - Central banks: 0.3 - Credit institutions: 0.5 - General governments: 1.8 - Non-financial corporations: 0.4 - Montenegro — Grand Total: 0.0 - North Macedonia — Grand Total: 0.0 - Poland — Grand Total: 0.9 - Credit institutions: 0.3 - General governments: 0.4 - Non-financial corporations: 0.2 - Romania — Grand Total: 2.7 - Central banks: 0.3 - General governments: 1.7 - Non-financial corporations: 0.5 - Serbia — Grand Total: 0.5 - Central banks: 0.2 - General governments: 0.2 - Non-financial corporations: 0.1 - Slovakia — Grand Total: 2.0 - Central banks: 0.1 - Credit institutions: 0.1 - General governments: 1.3 - Non-financial corporations: 0.5 - Slovenia — Grand Total: 0.2 - General governments: 0.2 - Ukraine — Grand Total: 0.2 - Central banks: 0.1 - General governments: 0.1 - Aggregate sector totals (Grand Total row, percentage of total assets): - Central banks: 4.7 - Credit institutions: 1.1 - Financial corporations other than credit institutions and investment firms: 0.3 - General governments: 9.6 - Households: 0.0 - Investment firms: 0.0 - Non-financial corporations: 2.8 - Grand Total (all sectors): 18.5 ### Implications and policy recommendations - Financial stability implication: losses originating in CESEE would intensify contagion via the most contagious banks and through interbank linkages. - Policy recommendations included in the source: - Adjust bank-wise systemic risk buffers as needed to counteract exposure to CESEE spillovers. - Carefully monitor and mitigate sanctions-related risks. - Monitor cross-border financial sector risks. ### Scenario analysis, PD elasticities, and interest-rate shocks (high-level) - PD elasticities for NFCs and households across Germany, Hungary, Czech Republic, Slovakia, Romania, Croatia, Croatia, and Romania are presented in multiple scenarios (Baseline scenario, Geopolitical scenario, Recession scenario) as multiples of 2025 values (figures shown in source). - Shocks to deposit and lending rates for activities in Hungary, Germany, Czech Republic, Slovakia, Romania, and Croatia are reported across the Baseline, Geopolitical, and Recession scenarios (annual shocks, in percentage points), illustrating heterogeneous interest-rate impacts by country and instrument. - Solvency stress-test outputs for SIs and LSIs show projected paths for: - SIs: Net Interest Margin (Percent of assets) under Baseline, Geopolitical, Recessionary scenarios for 2025–2028. - LSIs: Net Interest Margin under the same scenarios for 2025–2028. - Credit provisions and market risk gains or losses for SIs and LSIs in percent metrics (sources: IMF staff calculations). *Source: IMF staff calculations.* ### Appendix I. Econometric Models for Interest Rates on Deposits ### Appendix I. Econometric Models for Interest Rates on Deposits and on New Loans ### Models: Austria — Deposit rates (selected coefficients and fit) - Dependent variable: Change in Deposit Rate (Other sectors; Sight; Term; Sight; Term) - Lagged dependent variable: 0.200**, -0.378***, -0.0194, -0.403*** - D Euribor (1 month): 0.0307, 0.792***, 0.958*** - D Euribor (1 month) (t-1): 0.150***, 0.225*** - D 10Y bond yield: 0.114* - D 10Y bond yield (t-1): 0.0985 - Observations: 3,160; 3,102; 3,172; 1,884 - R2: 0.399; 0.417; 0.261; 0.360 - Robust standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1 - Regressions include bank fixed effects ### Models: Austria — Lending rates: Households and Retail (selected coefficients and fit) - Dependent variable: Change in interest rate (Retail: Consumer, Mortgage; Consumer, Mortgage) - Lagged dependent variable: -0.434***, -0.434***, -0.464***, -0.444*** - D Euribor (1 month): 0.526***, 0.549***, 1.080***, 1.290*** - D 10Y bond yield (t-1): 0.524***, 0.482***, 0.222*, 0.0760 - YoY Real GDP growth: 2.861**, 0.654, 0.115, 0.151 - Constant: 0.0362, -0.0471, 0.0191, -0.0218 - Observations: 11,822; 11,558; 17,569; 15,169 - R2: 0.241; 0.330; 0.301; 0.420 - Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 - Regressions include bank fixed effects - Fixed rate and Floating rate specifications reported ### Models: Austria — Lending rates: Other (NFCs) - (Table headings present but specific cell values not repeated here beyond other country tables) ### Models: Germany — Deposit rates and Lending rates (selected coefficients and fit) - Deposit rates (Change in interest rate; SME loans; Loans to Large Corporations) - Lagged dependent variable: -0.401***, -0.452*** - D Euribor (1 month): 0.639***, 0.434*** - D 10Y bond yield: 0.202, 0.302*** - Observations: 4,018; 1,600 - R2: 0.222; 0.301 - Lending rates (Change in interest rate; SME loans; Loans to Large Corporations) - Lagged dependent variable: -0.499***, -0.497*** - D Euribor (1 month): 0.968***, 1.080*** - D Euribor (1 month) (t-1): 0.387***, 0.222** - D 10Y bond yield (t-1): 0.114, 0.141 - YoY Real GDP growth (t-1): 0.446, 0.183 - Constant: 0.00289, 0.0232 - Observations: 8,177; 5,291 - R2: 0.387; 0.365 - Robust standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1 - Regressions include bank fixed effects - Fixed rate and Floating rate specifications reported ### Models: Country panels — Deposit and lending rate sensitivities (selected highlights) - Austria (short panel cells) - Change in 1 month euribor (t): 0.116**, 0.734***, 0.555***, 0.293***, 0.814***, 0.551*** - Change in 1 month euribor (t-1): 0.101** - Change in 10 year sovereign yield (t): 0.109, 0.211**, 0.0289, 0.212*, 0.565*** - YoY real GDP growth (t): 0.0326, 0.581 - Observations: 89 (various columns), R2: 0.744; 0.789; 0.608; 0.579; 0.791; 0.565 - Austria — New loan rates (mortgages and NFC loans) - Dependent variable (t-1): -0.315*** - Change in 1 month euribor (t): 0.171***, 0.660***, 0.465*** - Change in 10 year sovereign yield (t): 0.601***, 0.274**, 0.488*** - YoY real GDP growth: 0.522 - Observations: 89, 89, 88; R2: 0.872; 0.708; 0.465 - Czech Republic (selected coefficients) - Deposit rates: Dependent variable (t-1): -0.445*** - Change in 1 month euribor (t): 0.111***, 0.765***, 0.127***, 0.798***, 0.864*** - Change in 1 month euribor (t-1): 0.415*** - Change in 10 year sovereign yield (t-1): 0.0464**, 0.127, 0.122**, 0.108 - Observations: 85; 85; 69; 85; 85; 57 - R2: 0.511; 0.705; 0.125; 0.339; 0.666; 0.283 - Lending rates — mortgages and NFC loans: Dependent variable (t-1): -0.327* - Change in 1 month euribor (t): 0.211***, 0.761***, 0.532** - Change in 10 year sovereign yield (t): 0.103**, 0.773***, 0.227 - Observations: 85; 130; 30? (as printed: 85 31 30); R2: 0.664; 0.834; 0.433 - Slovakia (selected coefficients) - Deposit rates: Dependent variable (t-1): -0.414*** - Change in 1 month euribor (t): 0.0419***, 0.661***, 0.403***, 0.165***, 0.689***, 0.833*** - Observations: 69 (mostly), R2 range: 0.200; 0.526; 0.393; 0.296; 0.545; 0.315 - Lending rates — mortgages and NFC loans: - Dependent variable (t-1): -0.422***, -0.346** - Change in 1 month euribor (t): 0.170***, 0.527*, -0.112 - Change in 10 year sovereign yield (t): 0.203***, 0.728, 0.269 - Observations: 69, 59, 68; R2: 0.563; 0.341; 0.200 - Hungary (selected coefficients) - Deposit rates: Dependent variable (t-1): 0.150* - Change in 1 month euribor (t): 0.0544***, 0.852***, 0.479***, 0.174***, 0.877***, 0.562* - Observations: 81; 82; 82; 82; 82; 78 - R2: 0.473; 0.873; 0.386; 0.680; 0.834; 0.252 - Lending rates: - Dependent variable (t-1): -0.518** - Change in 1 month euribor (t): 0.153*, 0.808***, 0.224 - Change in 10 year sovereign yield (t): 0.301***, 0.433**, 0.637 - Observations: 82; 31; 30; R2: 0.623; 0.916; 0.410 - Romania (selected coefficients) - Deposit rates: Dependent variable (t-1): 0.398***, 0.202***, -0.255, -0.480*** - Change in 1 month euribor (t): 0.161***, 0.601***, 0.162, 0.172***, 0.823*** - Change in 1 month euribor (t-1): 0.0976**, 0.256***, 0.319** - Observations: 72; 72; 72; 73; 73; 69 - R2: 0.751; 0.859; 0.284; 0.523; 0.840; 0.329 - Lending rates — mortgages and NFC loans: - Dependent variable (t-1): 0.361***, -0.366**, -0.384* - Change in 1 month euribor (t): 0.627***, 0.610***, 0.615 - Change in 10 year sovereign yield (t): 0.615***, 0.203*, 0.249**, 0.111 - Change in 10 year sovereign yield (t-1): 0.430***, 0.294***, 0.911** - Observations: 72; 72; 31; 30; R2: 0.389; 0.826; 0.909; 0.660 - Croatia (selected coefficients) - Deposit rates: Dependent variable (t-1): -0.315**, 0.247**, 0.120, -0.323** - Change in 1 month euribor (t): 0.00322, 0.187, 0.0101, 5.12e-05, 0.140, 0.0222 - Change in 1 month euribor (t-1): 0.157, 0.101**, 0.338*** - Observations: 54; 53; 53; 54; 53; 53 - R2: 0.087; 0.233; 0.244; 0.14; 0.510; 0.117 - Lending rates: - Dependent variable (t-1): -0.315**, -0.218, -0.330** - Change in 1 month euribor (t): 0.112, 0.374**, 0.423***, 0.438 - Observations: 54; 53; 53; 50; R2: 0.266; 0.363; 0.355; 0.378 - Notes: Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1. Dependent variables reported vary by table (overnight deposit, term deposit up to 1 year, term deposit 1–2 years, interest rates on new mortgages, interest rates on new NFC loans up to 1 year, 1–5 years). Many regressions include bank fixed effects. ### Box 7. Austria: Corporate Credit Risk Model — Methodology and mapping - Scope and baseline: - Forward‑looking financial statements generated for almost 23,000 listed non‑financial corporations in 42 countries. - 2024 balance‑sheet data used as a common anchor. - Projection approach: - Country‑specific panel regressions estimate key performance ratios—sales growth, return on assets, leverage and the interest‑coverage ratio—based on their own lags and two exogenous drivers: real‑GDP growth and a broad financial‑conditions index that captures changes in spreads and equity valuations. - Macro inputs are taken directly from the IMF’s World Economic Outlook baseline and the FSAP adverse scenarios; each firm’s projected path is conditioned on the relevant national macro narrative. - Accounting identities convert projected income‑statement items into stock variables—debt, equity, cash buffers—over a three‑year horizon. - Interest expenses evolve with the firm’s historical funding mix and scenario‑specific shifts in short‑ and long‑term corporate borrowing rates. - Interest rate shocks of the current year applied to portion of debt maturing within the year; interest rate shock of year +1 applied to portion of debt maturing after one year. - Maturity structure (remaining maturity of less than one year, above one year) assumed unchanged during scenarios. - Mapping to PDs: - Simulated balance‑sheet indicators mapped into one‑year forward PDs using a matrix originally built by Moody’s for U.S. data over 1970‑2012 that assigns empirical default frequency to cells defined by interest‑coverage ratio bands and debt‑to‑equity bands. - Each firm placed in a cell for each projected year; corresponding historical default rate becomes its PD. - Aggregation across firms with debt weights yields a scenario‑specific corporate default rate for each country. - Benchmarking and calibration: - Because the Moody’s matrix is U.S.‑centric and reflects an earlier credit cycle, aggregated PDs are benchmarked to country‑specific default measures. - Two alternative anchors: - For a market‑based view of large‑listed corporates, the country aggregate for 2025 is rescaled to match Moody’s‑KMV one‑year expected default frequency (EDF); both mean and median EDFs are considered. - The resulting scaling factor is applied to the entire forward path of raw PDs under each macro scenario so differences across scenarios reflect only macro shocks and not cross‑country calibration quirks. *Appendix I. Econometric Models for Interest Rates on Deposits and on New Loans — IMF* ### Appendix II. Liquidity Stress Test Parameters ### Appendix II. Liquidity Stress Test Parameters ### HQLA Parameters (COREP 72.00) - Panel A — TOTAL UNADJUSTED LIQUID ASSETS (Scenario 1 / Scenario 2) - Coins and banknotes: 100.00% / 100.00% - Withdrawable central bank reserves: 100.00% / 100.00% - Central bank assets: 100.00% / 100.00% - Central government assets: 87.00% / 81.00% - Regional government / local authorities assets: 87.00% / 81.00% - Public Sector Entity assets: 87.00% / 81.00% - Recognisable domestic and foreign currency central government and central bank assets: 87.00% / 81.00% - Credit institution (protected by Member State government, promotional lender) assets: 87.00% / 81.00% - Multilateral development bank and international organisations assets: 100.00% / 100.00% - Qualifying CIU shares/units: underlying is coins/banknotes and/or central bank exposure: 100.00% / 100.00% - Qualifying CIU shares/units: underlying is Level 1 assets excluding extremely high quality covered bonds: 95.00% / 95.00% - Alternative Liquidity Approaches: Central bank credit facility: 100.00% / 100.00% - Central institutions: Level 1 assets excl. EHQ CB which are considered liquid assets for the depositing credit: 0.00% / 0.00% - Alternative Liquidity Approaches: Level 2A assets recognised as Level 1: 80.00% / 80.00% - Extremely high quality covered bonds (Level 1 EHQ): 93.00% / 93.00% - Qualifying CIU shares/units: underlying is extremely high quality covered bonds: 88.00% / 88.00% - Central institutions: Level 1 EHQ covered bonds which are considered liquid assets for the depositing credit: 100.00% / 100.00% - Total unadjusted level 2 assets components (selected) - Regional government / local authorities or Public Sector Entity assets (Member State, RW20%): 85.00% / 85.00% - Central bank or central / regional government or local authorities or Public Sector Entity assets (Third Country): 85.00% / 85.00% - High quality covered bonds (CQS2): 85.00% / 80.00% - High quality covered bonds (Third Country, CQS1): 85.00% / 80.00% - Corporate debt securities (CQS1): 85.00% / 80.00% - Qualifying CIU shares/units: underlying is Level 2A assets: 80.00% / 80.00% - Total unadjusted LEVEL 2B assets components (selected) - Asset-backed securities (residential, CQS1): 75.00% / 75.00% - High quality covered bonds (RW35%): 70.00% / 70.00% - Corporate debt securities (CQS2/3): 50.00% / 50.00% - Shares (major stock index): 50.00% / 50.00% - Restricted-use central bank committed liquidity facilities: 100.00% / 100.00% - Qualifying CIU: underlying is corporate debt securities (CQS2/3) shares or non-interest bearing assets (CQS3-5): 45.00% / 45.00% - Deposits by network member with central institution (no obligated investment): 75.00% / 75.00% ### Outflow Parameters (COREP 73.00) — Selected Items - Retail deposits - deposits exempted from the calculation of outflows: 0.00% / 0.00% - deposits where the payout has been agreed within the following 30 days: 100.00% / 100.00% - deposits subject to higher outflows — category 1: 14.9% / 14.9% - deposits subject to higher outflows — category 2: 19.9% / 19.9% - stable deposits: 5.00% / 5.00% - derogated stable deposits: 3.00% / 3.00% - other retail deposits: 10.00% / 10.00% - Operational deposits - maintained for clearing, custody, cash management or other comparable services covered by DGS: 5.00% / 5.00% - maintained for clearing, custody, cash management or other comparable services not covered by DGS: 25.00% / 25.00% - maintained in the context of IPS or a cooperative network — not treated as liquid assets for the depositing institution: 25.00% / 25.00% - maintained in the context of IPS or cooperative network — treated as liquid assets for the depositing credit institution: 100.00% / 100.00% - maintained in the context of an established operational relationship (other) with non-fin customers: 25.00% / 25.00% - maintained to obtain cash clearing and central credit institution services within a network: 25.00% / 25.00% - Excess operational deposits - deposits by financial customers: 100.00% / 100.00% - deposits by other customers covered by DGS: 20.00% / 20.00% - deposits by other customers not covered by DGS: 40.00% / 40.00% - Non-operational deposits - correspondent banking and provisions of prime brokerage deposits: 100.00% / 100.00% - deposits by financial customers: 100.00% / 100.00% - deposits by other customers covered by DGS: 20.00% / 20.00% - deposits by other customers not covered by DGS: 40.00% / 40.00% - Additional outflows (selected) - collateral other than Level 1 assets collateral posted for derivatives: 20.00% / 20.00% - Level 1 EHQ Covered Bonds assets collateral posted for derivatives: 10.00% / 10.00% - material outflows due to deterioration of own credit quality: 100.00% / 100.00% - impact of an adverse market scenario on derivatives transactions: 100.00% / 100.00% - outflows from derivatives: 100.00% / 100.00% - short positions — covered by collateralized SFT: 0.00% / 0.00% - short positions — other: 100.00% / 100.00% - callable excess collateral: 100.00% / 100.00% - internal netting of client’s positions: 50.00% / 50.00% - Committed facilities (selected) - credit facilities to retail customers: 5.00% / 5.00% - credit facilities to non-financial customers other than retail customers: 10.00% / 10.00% - credit facilities to credit institutions for funding promotional loans of retail customers: 5.00% / 5.00% - credit facilities to credit institutions for funding promotional loans of non-financial customers: 10.00% / 10.00% - other credit facilities to credit institutions: 40.00% / 40.00% - within IPS or cooperative network if treated as liquid asset by the depositing institution: 75.00% / 75.00% - liquidity facilities to non-financial customers other than retail customers: 30.00% / 30.00% - liquidity facilities to personal investment companies: 40.00% / 40.00% - liquidity facilities to SSPEs to purchase assets other than securities from non-financial customers: 10.00% / 10.00% - liquidity facilities to other financial customers: 100.00% / 100.00% - Other products and services — Scenario differences (Scenario 1 / Scenario 2) - Uncommitted funding facilities: 25.00% / 40.00% - undrawn loans and advances to wholesale counterparties: 25.00% / 40.00% - mortgages agreed but not yet drawn down: 25.00% / 40.00% - credit cards: 25.00% / 40.00% - overdrafts: 25.00% / 40.00% - planned outflows related to renewal or extension of new retail or wholesale loans: 25.00% / 40.00% - derivatives payables: 25.00% / 40.00% - trade finance off-balance sheet related products: 5.00% / 5.00% - others: 25.00% / 40.00% - Other liabilities and due commitments (selected) - liabilities resulting from operating expenses: 0.00% / 0.00% - liabilities in the form of debt securities if not treated as retail deposits: 100.00% / 100.00% - the excess of funding to retail customers: 100.00% / 100.00% - assets borrowed on an unsecured basis: 100.00% / 100.00% ### Outflows from Secured Lending and Capital Market-Driven Transactions - Counterparty is central bank (selected) - level 1 excl. EHQ Covered Bonds collateral: 5.00% / 5.00% - level 1 EHQ Covered Bonds collateral: 5.00% / 5.00% - level 2A collateral: 5.00% / 5.00% - level 2B asset-backed securities (residential or automobile, CQS1) collateral: 5.00% / 5.00% - Counterparty is non-central bank (selected) - level 1 excl. EHQ Covered Bonds collateral: 5.00% / 5.00% - level 1 EHQ Covered Bonds collateral: 15.00% / 15.00% - level 2A collateral: 20.00% / 20.00% - level 2B asset-backed securities (residential or automobile, CQS1) collateral: 25.00% / 25.00% - level 2B covered bonds: 40.00% / 40.00% - other Level 2B assets collateral: 60.00% / 60.00% - non-liquid assets collateral: 100.00% / 100.00% ### Cash-flow Stress Test Parameters (COREP 66.01) - Panel B — Open Maturity Parameters for Calibrating 12-Month Cumulative Outflows (selected) - Liabilities resulting from securities issued (if not treated as retail deposits) — unsecured bonds due: 50% / 100% - regulated covered bonds: 50% / 100% - securitisations due: 50% / 100% - Level 1 central bank: 100% / 100% - Level 1 (CQS 1): 100% / 100% - Level 2A corporate bonds (CQS1): 100% / 100% - Level 2B Asset Backed Securities (ABS) (CQS1): 100% / 100% - Liabilities not reported in 1.2, resulting from deposits received (excluding deposits received as collateral) — selected: - stable retail deposits: 19% / 0.16% and 5% / 0.17% (table shows term / openmat combinations) - other retail deposits: 44% / 0.44% and 10% / 0.39% - operational deposits: 15% / 0.51% and 49% / 2.32% - non-operational deposits from credit institutions: 50% / 0.52% and 100% / 100% - non-operational deposits from other financial customers: 50% / 0.52% and 100% / 100% - non-operational deposits from central banks: 25% / 0.22% and 100% / 100% - non-operational deposits from non-financial corporates: 25% / 0.22% and 76% / 0.97% - non-operational deposits from other counterparties: 25% / 0.22% and 76% / 1.11% - FX-swaps maturing: 50% / 100% - Derivatives amount payables other than those reported in 1.4: 50% / 100% - Other outflows: 50% / 100% - Outflow risk decompositions (Gen risk aver / Idio) for selected items - stable retail deposits: 0.04% / 0.16% - other retail deposits: 0.08% / 0.44% - operational deposits: 0.26% / 0.26% - non-operational deposits from central banks: 0.22% / 0.22% - non-operational deposits from non-financial corporates: 1.07% / 1.07% - non-operational deposits from other counterparties: 1.07% / 1.07% ### Inflow Parameters and Counterbalancing Capacity (COREP 66.01) - Panel C — Inflow Parameters (selected) - Monies due from secured lending and capital market driven transactions collateralised by Level 1 tradable assets: 100% / 100% - Monies due not reported in 2.1 resulting from loans and advances granted to: - retail customers: 2% / 0% and 2% / 0% (term / openmat combinations) - non-financial corporates: 2% / 0% and 2% / 0% - credit institutions: 100% / 100% and 100% / 100% - other financial customers: 100% / 100% and 100% / 100% - central banks: 100% / 100% and 100% / 100% - other counterparties: 3% / 0% and 3% / 0% - FX-swaps maturing (inflows): 100% / 100% - Derivatives amount receivables other than those reported in 2.3: 100% / 100% - Paper in own portfolio maturing: 100% / 100% - Other inflows: 100% / 100% - Panel D — Counterbalancing Capacity and Contingency Parameters (selected) - Counterbalancing capacity — coins and bank notes: 100% / 100% - Withdrawable central bank reserves: 100% / 100% - Level 1 tradable assets — Level 1 (CQS 1): 98% / 98% - Level 1 tradable assets — Level 1 (CQS2, CQS3): 95% / 95% - Level 1 tradable assets — Level 1 (CQS4+): 75% / 75% - Level 1 covered bonds (CQS1): 97% / 97% - Level 2A tradable assets — Level 2A corporate bonds (CQS1): 90% / 90% - Level 2B tradable assets — Level 2B ABS (CQS1): 75% / 75% - other tradable assets — central government (CQS1): 75% / 75% - non tradable assets eligible for central banks: 62% / 62% - own issuances eligible for central banks: 62% / 62% - undrawn committed facilities received — Level 1 facilities: 25% / 25% - undrawn committed facilities received — Level 2B restricted use facilities: 25% / 25% - Net change of Counterbalancing Capacity and Cumulated Counterbalancing Capacity (table entries present but no numeric totals shown) - Contingencies (selected) - Outflows from committed credit facilities considered as Level 2B by the receiver: 40% / 40% - Outflows from uncommitted funding facilities: 0% / 0% - Outflows due to downgrade triggers: 25% / 25% ### Banking Sector: Solvency Stress Test Matrix — Top-down by IMF (Table 15, A) - Institutional perimeter and data (selected) - Institutions included: 7 SI banks, plus Unicredit subsidiary in Austria, and 303 LSI banks - Market share: About 95 percent of the banking sector assets - Data and baseline date: Data vintage: 2024 Q4 for balance sheet data and P&L items, capital ratios and other supervisory data such as NII adjusted to 2025 Q3. - Supervisory data sources: Bank balance sheet and supervisory statistics (including FINREP and COREP), IRRBB, short-term exercise (STE), provided by the OeNB and the ECB. - Additional data sources: ITS templates, 2021 HFCS for household vulnerability analysis, MCM Corporate Stress Test based on Datastream and Capital IQ for Non-Financial Corporates analysis. - Scope of consolidation: banking activities at the consolidated level for banks having their headquarters in Austria, and Unicredit Austria subsidiary. - Coverage of sovereign and non-sovereign securities exposures: debt securities measured through fair value (FVPL and FVOCI) and amortized cost (AC) account. *Appendix II. Liquidity Stress Test Parameters (PDF chapter).* ### 2. Channels of Risk ### 2. Channels of Risk ### Propagation — Methodology - Derived from the 2025 EA FSAP and France FSAP methodology, enriched with Austrian FSAP specific models and drawing on the OeNB stress testing framework including reverse ownership income and dividend channels. - For internally modelled exposures (IRB): projection of PiT and TTC PDs, PiT and DT LGDs, EAD, and RWA. - For SA exposures: projection of new flows of defaulted exposures and RWA based on risk weights; performing and non-performing loans separately. - For SIs: provisioning for IRB and SA portfolios; for LSIs under national GAAP using standard NPL modeling. - Balance-sheet regulatory approach. - Provisioning for IRB and SA are modeled using an expected loss approach. - Traded risk impact from revaluation of trading assets (FVPL) and securities classified as FVOCI or AFS assessed using a modified duration approach or sensitivities to market risk factors (Greeks) with hedging strategy considered based on ITS and STE data if available (based on the 2025 EA FSAP approach for SIs). - Time-to-repricing approach for Interest income and expense from IRRBB J-template. - Inclusion of a participation risk channel, reflecting dividend reductions and a write-down of participation book values for banks with equity stakes in other banks, based on OeNB’s methodology. Satellite models for macro‑financial linkages - Models for credit losses, funding costs, lending rates, and assumptions for other P&L components. - Household and corporate PD analysis using micro-data at individual household (based on household survey HFCS) and corporate (based on commercial corporate database). - Credit risk from Credit register data considered for sectoral shocks of domestic exposures; default probabilities specific to Austria’s real estate sectors are estimated. - Credit risk models for cross-border exposures based on the household survey model and a model based on aggregate EDFs. - LGDs for collateralized exposures linked to paths for real estate prices in the scenario using a smoothing factor to account for the TTC regulatory approach. - Interest income projected at portfolio level and country of exposure using a structural approach applying interest rate shocks derived from satellite models on new originations and loans’ repricing ladder to outstanding volumes. - Funding costs projected at portfolio level using funding structure by counterpart (retail, wholesale) and maturity bucket (overnight vs. term), and by product (deposit, borrowings), and country of exposure. ### Tail Shocks — Stress test horizon and scenarios - Stress test horizon: 2026 Q1– 2028 Q4 (three years). - Scenarios (for Austria and main cross-border exposures: Germany and four CESEE countries): - Baseline scenario updated from the Spring 2026 WEO macroeconomic projections. - Adverse scenario 1 derived from the EA FSAP: A geopolitical scenario (or higher-for-longer) featuring an escalation of geopolitical conflicts. - Adverse scenario 2 derived from the EA FSAP: A recessionary scenario showing a synchronized global slowdown amplified by sovereign debt distress in the EA. - The two adverse scenarios rely on GFM, a structural macro-econometric model of the world economy, disaggregated into forty national economies, documented in Vitek (2018), except for Hungary, Slovakia, Romania and Croatia for which a BVAR model was developed to simulate macro variables under the adverse scenarios. Second-round effects and sensitivity analysis - Second round effects of recapitalizations to Institutional Protection Schemes (IPS) for solvency. - Household borrower-based instruments and mortgage default – counterfactual analysis. - Dividend and market losses on participations. - Solvency and liquidity risk interactions, mainly through fire sales and funding costs and based on 2023 GFSR methodology. ### Risks and Buffers - Risks covered: credit (on loans to households and non-financial corporates, floating and fixed rate instruments for Austrian exposures), market (the P&L impact of net open positions in market risk factors such as foreign exchange risks) and interest rate risk on the banking book (IRRBB). - Domestic exposures and main cross-border exposures: Germany, Hungary, Czech Republic, Slovakia, Romania and Croatia. Behavioral adjustment - Static approach for balance sheet growth over the stress-test horizon: asset allocation and funding composition remain the same; balance sheet remains static in nominal Euro terms. - No cures in the adverse scenarios. No write-offs. - RWAs projection: - Standardized portfolios: RWAs change due to balance sheet growth, new inflows of non-performing loans, exchange rate movements, and conversion of a portion of off-balance sheet items (undisbursed credit lines and guarantees) to on-balance sheet items. - IRB portfolios: through-the-cycle-PDs, downturn LGDs and EAD for each asset class/industry used to project risk weights. - Interest income from nonperforming loans is not accrued. - Dividends are paid out by banks that remain profitable and adequately capitalized throughout the stress. Dividend rate: average ratio between observed dividends and profits after tax over recent years. - Tax rate aligned with 2025 EBA methodology. - Model includes dividend payments and equity valuation losses from reverse ownership structure of the decentralized sector. ### Regulatory and Market-Based Standards and Parameters - National regulatory framework: Basel III regulatory minima on CET1 (4.5 percent) plus Pillar II requirements (total SREP capital requirement), and include any requirements due to systemic buffers (SyRB, O-SII buffer), and capital conservation buffer. - Leverage ratio during the stress test horizon against the 3 percent Basel III minimum requirement. ### Reporting Form for Results - Output presentation: - Capital path under various scenarios by groups of banks, aggregate and categorized by type of bank (SIs, LSIs). - System-wide capital shortfall. - Number of banks and percentage of banking assets in the system that fall below regulatory minima or breach capital buffers. - Outputs also include information on impact of different result drivers, including profit components. ### B. Banking Sector: Liquidity Stress Test (Top-Down by IMF) Institutional perimeter and data - Institutions included: 7 SI banks, plus UniCredit subsidiary in Austria, and 303 LSI banks. - Market share: More than 95 percent of the deposit-taking institutions. - Data vintage and time span: monthly from October 2020 to September 2025 for SIs. September 2025 for LSIs. - Data sources: Supervisory data (FINREP, COREP), MFI interest rates data from the national authorities and the ECB. - Scope of consolidation: Consolidated level. Methodology and horizon - LCR-based tests, using regulatory parameters and more severe scenarios; breakdown by aggregate currency positions for significant currencies (CZK and EUR). - Cash flow-based liquidity stress test focusing on net liquidity position (difference between aggregate cumulated CBC and net outflows); breakdown by aggregate currency positions for significant currencies (CZK and EUR). - Further liquidity stress test incorporating second-round effects from IPS for liquidity reallocation between banks and use of pre-contributed liquidity pools in response to funding gaps in the LCR-based stress tests and the cash flow stress tests. - Stress test horizon: 30 days for LCR-based tests; from overnight up to 12 months with varying time buckets for cash flow-based tests. Type of analyses - Stress scenarios derived from similar scenarios in the 2025 Euro Area FSAP, with varying intensity of adverse liquidity conditions affecting deposit outflows and asset reevaluations. - Both general risk aversion shock and idiosyncratic shocks to individual banks are examined. - Deposit runoff rates and outflow rates for several unsecured outflow items re-calibrated to Austrian-specific conditions. - Reverse stress tests. Buffers and regulatory standards - Liquidity from the central bank is not considered. - Buffers: capacity of banks to generate liquidity from inflows and from assets under stress (counterbalancing capacity). - For LCR-based analysis, the hurdle rate is set at 100 percent at the aggregate currency level (per Basel III and domestic regulation). - For cash flow-based analysis, outcomes of interest are the net liquidity position and the survival period. - Both analyses examine results before and after the use of IPS support mechanisms. Reporting format for results - Output includes: (1) Average LCR, Net Liquidity Position, and survival period; (2) Market share of institutions with LCR below regulatory limits and with negative net liquidity position. ### C. Banking Sector: Contagion and Interconnectedness Analysis (Top Down by IMF) Institutional perimeter and data - Institutions included: 7 SI banks, plus UniCredit subsidiary in Austria, and up to 325 LSI banks. - Market share: More than 95 percent of the deposit-taking institutions. - Data vintage: December 2024. - Data: Gross interbank exposures including participation instruments. - Data source: OeNB. - Scope of consolidation: Consolidated level. Methodology and horizon - Model calibrated for the Austrian banking system (CoMap: Mapping Contagion in the Euro Area Banking Sector). - IPS mechanism modelled through LGDs: for the decentralized sector individual (Tier 1) banks assumed to be 0.9 (no IPS scenarios) or actual data (with IPS scenarios). - Stress test horizon: 1 month. Type of analyses - Scenario analysis: - “Baseline scenario”: IPS in place for the decentralized sector (lower LGDs for Tier 1 decentralized sector banks), with central bank support, both credit and liquidity channel in place. - “No IPS” scenario: (1) amended by higher LGDs for Tier 1 decentralized sector banks. - “CESEE loss” scenario: shocking capital to the nonfinancial corporate large CESEE exposures multiplied by 5 times the PDs in ARNIEs. Buffers and behavioral adjustments - Approach simulates cascading defaults and counterparty credit losses; counterfactual modelling of lack of IPS in the decentralized sector. - Buffers: CET1 over bank specific regulatory leverage ratio (solvency), and counterbalancing capacity minus non tradable assets eligible for central banks minus HQLA. Regulatory standards - Bank specific minimum CET1 ratio (Pillar 1+Pillar 2R) (solvency), and liquidity reserves (HQLA) minimum net outflows (liquidity). Reporting format for results - Outputs include: (1) Capital depleted due to triggered defaults; (2) Normalized Contagion Index; and (3) Normalized Vulnerability Index. ### D. Banking Sector: Analysis on Large Exposures to CESEE (Top Down by IMF) Institutional perimeter and data - Institutions included: 7 SI banks, plus UniCredit subsidiary in Austria. - Market share: 68 percent of the deposit-taking institutions. - Data vintage: December 2024. - Data: Large exposures to CESEE countries with and without credit mitigation. - Data source: COREP, FINREP. - Scope of consolidation: Consolidated level. Methodology and analysis - Methodology: Reporting large exposures data. - Type of analyses: Data reporting and scenario analysis in Contagion and Interconnection Analysis, analyzing large exposures by country, by counterparty (General government, non-financial corporations, households, other financial institutions). Reporting format for results - Output presentation: Table without credit mitigation, by country, by counterparty. Note - The authorities recommended that the stress tests not rely on intra-year data which may have important seasonality aspects, especially for LSIs. *Source: 2. Channels of Risk (Austria FSAP chapter).* --- _Source: https://www.imf.org/-/media/files/publications/cr/2026/english/1autea2026007.pdf_