## wpiea2026094-source-pdf

## Source details

**Canonical URL:** [wpiea2026094-source-pdf](https://www.imf.org/-/media/files/publications/wp/2026/english/wpiea2026094-source-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2026/english/wpiea2026094-source-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2026/english/wpiea2026094-source-pdf.pdf.json)

---

### I. Introduction and scope
- Purpose and scope
  - Update of the Laeven and Valencia database to cover crisis episodes during 1970–2025 using an integrated narrative (textual) and quantitative approach.
  - Database records 164 banking crises (including borderline cases), dated yearly and, where possible, monthly.
  - Coverage includes policy responses, fiscal costs, output losses, and other statistics across countries of all income levels.
- Context and motivation
  - Since 2020, overlapping shocks (COVID-19 shutdown, supply chain disruptions, global inflation surge) increased macroeconomic volatility and elevated debt burdens; widespread systemic banking crises did not materialize mainly due to improved bank capitalization, financial regulation, and prompt policy responses.
  - Banking vulnerabilities resurfaced in 2023 (failures of SVB, Signature Bank, First Republic in the US; emergency takeover of Credit Suisse in Switzerland), which did not meet the paper’s definition of systemic banking crises but highlighted risks (interest rate risk, depositor flight, crisis management).

### II. Definition and operational thresholds of a banking crisis
- Two required conditions for a banking crisis (Laeven and Valencia 2013, 2020):
  - 1) Significant signs of financial distress in the banking system (significant bank runs, losses, and/or bank liquidations).
  - 2) Significant banking policy intervention measures in response to significant losses.
- Crisis dating rule
  - The first year in which both criteria are met is treated as the year when the crisis became systemic.
- Severe-loss sufficient condition (operationalization)
  - Losses are considered severe when either:
    - (i) a country’s banking system exhibits significant losses resulting in a share of nonperforming loans above 20 percent of total loans or bank closures of at least 20 percent of banking system assets; or
    - (ii) fiscal restructuring costs of the banking sector exceed 5 percent of GDP.
  - Note: the 20 percent NPL threshold requires that the ratio at least doubled in three years (footnote).
- Significance of policy intervention (must meet at least three of six measures)
  - 1) deposit freezes and/or bank holidays;
  - 2) significant bank nationalizations;
  - 3) bank restructuring fiscal costs (at least 3 percent of GDP);
  - 4) extensive liquidity support (at least 5 percent of deposits and liabilities to nonresidents);
  - 5) significant guarantees put in place;
  - 6) significant asset purchases (at least 5 percent of GDP).
- Precise definitions of policy variables (verbatim operationalizations included for):
  - Deposit freeze and bank holidays, Significant nationalizations, Significant bank guarantees, Liquidity support, Bank restructuring costs, Asset purchases.

### III. Identification and textual-analysis methodology
- Data sources and extraction
  - Primary textual source: publicly available IMF country reports (Article IV staff reports, technical assistance reports, financial sector stability assessments) published between 2017 and 2025.
  - Supplementary sources: international news articles (Factiva), official communications from national central banks (press releases), national authorities’ technical reports (e.g., central bank financial stability reports).
  - Raw text extracted using PyPDF2; reports categorized by country and year with a Python script.
- Textual analysis pipeline
  - Constructed a dictionary of about 250 unique words and expressions indicative of banking stress.
  - Matched extracted words against the dictionary to short-list potential crisis episodes based on frequency of “banking crisis” terms.
  - Short-listed IMF reports and supplementary materials were manually reviewed to verify events and assess whether the crisis definition was met.
- Integration with expert judgment
  - Final identification and dating combined textual signals with manual expert evaluation and supporting quantitative evidence.
- Text preprocessing and normalization (implementation)
  - Steps: splitting into sentences; removing punctuation and unwanted characters; eliminating section headers, titles, table rows, hyperlinks, and footnotes.
  - Sentence-level processing: tokenization, lemmatization, lowercase conversion, stopword removal, retaining only valid English words.
  - Tools: WordNet lemmatizer from NLTK in Python; lemmatized words stored in the dataset.
- Dictionary examples (selected entries verbatim)
  - Examples include terms such as: bank run, bank failure, deposit freeze, deposit insurance, liquidity support, asset purchase, recapitalization, non performing, npl, blanket guarantee, central bank intervention, emergency liquidity, government recapitalization, systemic risk, etc.

### IV. Aggregate update and newly identified episodes (1970–2025)
- Update summary
  - Identified 13 new crisis episodes since Laeven and Valencia (2020), bringing total to 164 systemic banking crises (including borderline episodes) for 1970–2025.
  - Identified 3 borderline episodes that did not meet the full definition: Nicaragua (2018), Vietnam (2022), and Sri Lanka (2023).
- Newly identified episodes (relative to Laeven and Valencia 2020)
  - Angola — 2015
  - Azerbaijan — 2015
  - Chad — 2015
  - Equatorial Guinea — 2015
  - Ghana — 2015
  - Kazakhstan — 2014
  - Lebanon — 2019
  - Nicaragua* — 2018
  - Republic of the Congo — 2018
  - São Tomé and Príncipe — 2016
  - Sri Lanka* — 2023
  - Tajikistan — 2016
  - Vietnam* — 2022
  - (* Denotes borderline cases)
- Character and distributional facts
  - New episodes predominantly triggered by significant negative shocks outside the financial sector (e.g., sharp decline in oil prices starting late 2014 affecting oil-producing countries such as Angola, Chad, Azerbaijan).
  - Most countries experienced at least one systemic banking crisis during 1970–2025; few countries experienced more than two episodes (including borderline cases): Argentina (4), Chad (3), Democratic Republic of Congo (3), Nicaragua (3), Ukraine (3).

### V. Special 2023 events: United States and Switzerland (assessment under defined thresholds)
- United States (March–May 2023)
  - Failures in rapid succession: Silicon Valley Bank (SVB), Signature Bank, First Republic.
  - Combined deposits of the three banks: US$ 367 billion, or 2.0 percent of total US bank deposits.
  - SVB and Signature Bank had about 90% uninsured deposits; bank runs occurred at unprecedented speed.
  - Federal Reserve actions: new liquidity facility; systemic risk exception invoked; uninsured depositors in SVB and Signature Bank were fully protected by the FDIC, backed by a US Treasury guarantee.
  - First Republic failed on May 1, 2023 (closed after downgrades and large-scale deposit withdrawals); FDIC sold the bank and almost all its assets to J.P. Morgan Chase.
  - Assessment vs. crisis definition:
    - Peak liquidity support amounted to about 1% of total US banks liabilities (below 5 percent threshold).
    - No fiscal costs from bank restructuring.
    - Government guarantees applied explicitly to affected institutions but not to the banking system.
  - Conclusion: did not meet the paper’s criteria for a systemic banking crisis.
- Switzerland (Credit Suisse)
  - Credit Suisse represented 14.2 percent of Swiss total banking sector assets.
  - Swiss National Bank injected 168 billion Swiss francs (around US$ 185 billion) in emergency liquidity support (around 20% of Swiss GDP).
  - Taken over by UBS with a 9 billion Swiss franc state guarantee on the value of the bank’s assets.
  - Liquidity support provided by SNB represented 4.5 percent of total banking system liabilities in Switzerland (below 5 percent threshold).
  - Aside from state guarantee on bank’s liabilities to support takeover, there were no fiscal outlays from bank restructuring costs.
  - Assessment: despite being a systemic institution, the event does not qualify as a systemic banking crisis under the paper’s definition due to speedy resolution and thresholds not met.

### VI. Policy interventions during crises — instruments and stylized facts
- Overview
  - Both high-income and low and middle-income countries heavily rely on liquidity support during crises.
  - High-income countries tend to deploy a wider array of instruments (e.g., coordinated central bank swap facilities) and make greater use of significant guarantees on bank liabilities.
  - Deposit freezes and bank holidays are rare but recent cases include deposit freezes: Cyprus (2013), Ukraine (2014), Greece (2015), Lebanon (2019); bank holidays: Cyprus, Greece, Lebanon, Sri Lanka.
- Stylized facts and key statistics
  - Peak liquidity support:
    - Liquidity support measured as ratio of central bank claims on financial sector to deposits and foreign liabilities.
    - Median peak liquidity support ratio in the sample reaches 20 percent.
  - Guarantees:
    - Significant guarantees on bank liabilities are more common among high-income countries.
  - Recapitalization and restructuring:
    - Bank recapitalization with public funds is often the most important component of direct fiscal costs.
    - Government capital injections often involve preferred and common equity and conditions (board seats, dividend restrictions); governments can retain ownership for many years.
- Examples from Table A2 (policy tools with exact reported values)
  - Iceland Sep-08: Asset Purchase 24.30 (percent of GDP); IMF Program 2008.
  - Ireland Sep-08: Bank Recapitalization 37.13 (percent of GDP); Asset Purchase Y; IMF Program 2010.
  - Greece 2008–2012: Bank Recapitalization 25.40 (percent of GDP); IMF Programs 2010, 2012.
  - Indonesia Nov-97 / Jan-98: Bank Recapitalization 37.30 (percent of GDP); duration 78 (months) for guarantees; Asset Purchase Y; IMF Program 1998.
  - Turkey Nov-00: Bank Recapitalization 24.50 (percent of GDP); Asset Purchase Y; IMF Program 2000.
  - United States Dec-07 / Oct-08: Bank Recapitalization 3.60 (percent of GDP); Asset Purchase Y.

### VII. Crisis outcomes: fiscal costs, peak NPLs, durations, and output losses
- Metrics and measurement rules
  - Direct fiscal costs: fiscal outlays linked to government interventions in the banking system, normalized by nominal GDP of the year of outlay.
  - Peak nonperforming loans (NPLs): peak NPL ratio in percent of total loans.
  - Crisis duration: number of years between start and end of crisis (end defined as the year before both real GDP growth and real credit growth are positive for at least two consecutive years); durations truncated at five years.
  - Output losses: 4-year cumulative deviation of output from pre-crisis trends (cumulative over [T, T+3], expressed as percent of one year’s trend GDP).
- A. Fiscal costs (summary statistics)
  - Median gross fiscal cost for high-income country crises: 6.7 percent of GDP (med = 6.67).
  - Median gross fiscal cost for low and middle-income country crises: 7.5 percent of GDP (med = 7.50).
  - Measurement notes: fiscal costs include capital injections, operating costs of asset management companies, exercised public guarantees that led to outlays; excludes government guarantees as not ex-post outlays and excludes emergency liquidity assistance.
  - Notable datum: For Lebanon (2019), direct fiscal costs are not available; 2023 Article IV consultation staff report noted estimated overall losses in the financial system at about US$70 bn (more than 300 percent of the estimated 2022 GDP).
- B. Peak Nonperforming Loans (NPLs)
  - Median peak NPLs:
    - High-income countries: med = 11.14 (peak NPL in percent of total loans).
    - Low and middle-income countries: med = 30.00 (peak NPL in percent of total loans).
  - Caveat: NPL measures affected by regulatory differences, forbearance, and definition heterogeneity.
- C. Systemic Banking Crises Duration
  - End-date rule: year before both real GDP growth and real credit growth are positive for at least two consecutive years; credit measured as bank credit to private sector (national currency) from IFS deflated by CPI from WEO; GDP from WEO; when credit data unavailable, end date determined by GDP growth rule.
  - Truncation: crisis durations truncated at five years to avoid upward bias.
  - Distribution:
    - About two-thirds of crises ended in less than five years.
    - Crises in high-income countries tend to be more persistent, often lasting five years or more.
    - Most crises in low and middle-income countries lasted four years or less.
- D. Output Losses (median cumulative values)
  - Median cumulative output loss:
    - High-income countries: med = 31.75 (cumulative output loss in percent of GDP).
    - Low and middle-income countries: med = 10.80 (cumulative output loss in percent of GDP).
  - Measurement specifics:
    - Trend computed via HP filter (λ=100 for annual data) on log real GDP over [T-20, T-1]; for crises starting in 2013 or later, calculations use real GDP data from the fall 2025 WEO; for earlier crises, output losses taken from Laeven and Valencia (2020).
  - Caveat: output losses reflect losses around crisis episodes but are not solely caused by banking crises; may include multiple concurrent shocks.

### VIII. Selected episode-level values (Table A1 examples; preserved precision)
- Albania 1994: Peak liquidity 7.6; Peak NPLs 26.8.
- Algeria 1990–1994: Output loss 5/; Fiscal costs 41.4; Peak liquidity 37.6; Peak NPLs 30.0.
- Argentina 2001–2003: Output loss 71.0; Fiscal costs 9.6; Peak liquidity 22.9; Peak NPLs 20.1.
- Austria 2008–2012: Fiscal costs 19.2; Output loss 5/; Peak liquidity 10.0; Peak NPLs 4.1.
- China 1998: Output loss 19.4; Fiscal costs 18.0; Peak liquidity 62.0; Peak NPLs 20.0.
- Greece 2008–2012: Output loss 64.9; Fiscal costs 28.7; Peak liquidity 61.7; Peak NPLs 37.1.
- Ireland 2008–2012: Output loss 107.7; Fiscal costs 37.6; Peak liquidity 18.1; Peak NPLs 25.7.
- United States 2007–2011: Output loss 30.0; Fiscal costs 4.5; Peak liquidity 4.7; Peak NPLs 5.0.
- Uruguay 2002–2005: Output loss 66.1; Fiscal costs 20.0; Peak liquidity 12.8; Peak NPLs 36.3.
- Vietnam 1997: Output loss 0.0; Fiscal costs 10.0; Peak liquidity 64.9; Peak NPLs 35.0.
- Note: durations truncated at 5 years where indicated (footnote 5); borderline cases flagged (footnote 6).

### IX. Conclusions, uses, and outlook
- Methodology and contribution
  - Textual analysis tools applied to IMF staff reports and supplementary sources improved detection of potential crisis episodes; expert evaluation and quantitative thresholds determine final inclusion and dating.
  - The update extends the Laeven and Valencia database through end-2025, enhancing comprehensiveness across income levels and documenting policy responses.
- Key findings
  - Systemic banking crises remained relatively rare between 2020 and 2025 despite major global shocks, reflecting increased resilience of the global banking sector and the impact of post-GFC reforms and policy support.
  - Banking turmoil in 2023 (US and Switzerland) underscores the continuing relevance of research on crisis avoidance, mitigation, and resolution even when events do not meet defined systemic thresholds.
- Uses and outlook
  - The database supports policy and analytical work: understanding cross-country vulnerability patterns, evaluating resolution frameworks, assessing macroeconomic effects of crises, and evaluating effectiveness of containment measures.

*Source: https://www.imf.org/-/media/files/publications/wp/2026/english/wpiea2026094-source-pdf.pdf*

### References .............................................................................................................

### wpiea2026094-source-pdf - References .....................................................................................................................

### References
- References .............................................................................................................................. 21

### Appendix I. Banking Crises Episodes and Policy Responses
- Appendix I. Banking Crises Episodes and Policy Responses ............................................. 25

*Source: https://www.imf.org/-/media/files/publications/wp/2026/english/wpiea2026094-source-pdf.pdf*

### Appendix II. Implementation of Textual Analysis ................................................................. 43

### Appendix II. Implementation of Textual Analysis

### I. Introduction
- Purpose and scope
  - Update of the Laeven and Valencia database to cover crisis episodes during 1970–2025 using an integrated narrative (textual) and quantitative approach.
  - Database records 164 banking crises (including borderline cases), dated yearly and, where possible, monthly.
  - Coverage includes policy responses, fiscal costs, output losses, and other statistics across countries of all income levels.
- Context and motivation
  - Since 2020, overlapping shocks (COVID-19 shutdown, supply chain disruptions, global inflation surge) increased macroeconomic volatility and elevated debt burdens; widespread systemic banking crises did not materialize mainly due to improved bank capitalization, financial regulation, and prompt policy responses.
  - Banking vulnerabilities resurfaced in 2023 (failures of SVB, Signature Bank, First Republic in the US; emergency takeover of Credit Suisse in Switzerland), which did not meet the paper’s definition of systemic banking crises but highlighted risks (interest rate risk, depositor flight, crisis management).

### II. Definition of a Banking Crisis
- Two required conditions for a banking crisis (Laeven and Valencia 2013, 2020):
  1. Significant signs of financial distress in the banking system (significant bank runs, losses, and/or bank liquidations).
  2. Significant banking policy intervention measures in response to significant losses.
- Crisis dating rule: the first year in which both criteria are met is treated as the year when the crisis became systemic.
- Severe-loss sufficient condition (operationalization):
  - Losses are considered severe when either:
    - (i) a country’s banking system exhibits significant losses resulting in a share of nonperforming loans above 20 percent of total loans or bank closures of at least 20 percent of banking system assets; or
    - (ii) fiscal restructuring costs of the banking sector exceed 5 percent of GDP.
  - Note: the 20 percent NPL threshold requires that the ratio at least doubled in three years (footnote).
- Significance of policy intervention: at least three out of six measures must be used:
  - 1) deposit freezes and/or bank holidays;
  - 2) significant bank nationalizations;
  - 3) bank restructuring fiscal costs (at least 3 percent of GDP);
  - 4) extensive liquidity support (at least 5 percent of deposits and liabilities to nonresidents);
  - 5) significant guarantees put in place;
  - 6) significant asset purchases (at least 5 percent of GDP).
- Precise definitions of policy variables (as used in the crisis definition):
  - Deposit freeze and bank holidays: whether restrictions on deposit withdrawals or bank holidays were introduced; duration and affected instruments collected when implemented.
  - Significant nationalizations: government takeover of systemically important financial institutions or majority stakes.
  - Significant bank guarantees: full protection of liabilities or extension to non-deposit liabilities; raising deposit insurance only is not included.
  - Liquidity support: central bank claims on other depository institutions (IMF IFS) and Treasury liquidity, normalized by total deposits and bank liabilities to nonresidents; considered extensive when ratio exceeds 5 percent and more than doubles relative to pre-crisis level.
  - Bank restructuring costs: gross fiscal outlays for restructuring (primarily recapitalizations); significant if exceeding 3 percent of GDP (excluding liquidity assistance from the treasury); gross costs reported (net reported if recoveries available).
  - Asset purchases: purchases of assets from financial institutions by central bank, treasury, or government entities; significant if exceeding 5 percent of GDP.

### III. Identifying and Dating Systemic Banking Crises
- Data sources and extraction
  - Primary textual source: publicly available IMF country reports (Article IV staff reports, technical assistance reports, financial sector stability assessments) published between 2017 and 2025.
  - Supplementary sources: international news articles (Factiva), official communications from national central banks (press releases), national authorities’ technical reports (e.g., central bank financial stability reports).
  - Raw text extracted using PyPDF2; reports categorized by country and year with a Python script.
- Textual analysis pipeline
  - Constructed a dictionary of about 250 unique words and expressions indicative of banking stress (Appendix II, Table A3).
  - Matched extracted words against the dictionary to short-list potential crisis episodes based on frequency of “banking crisis” terms.
  - Short-listed IMF reports and supplementary materials were manually reviewed to verify events and assess whether the crisis definition was met.
- Integration with expert judgment
  - Final identification and dating combined textual signals with manual expert evaluation and supporting quantitative evidence.

### IV. Banking Crises Episodes During 1970–2025
- Update summary
  - Identified 13 new crisis episodes since Laeven and Valencia (2020), bringing total to 164 systemic banking crises (including borderline episodes) for 1970–2025.
  - Identified 3 borderline episodes that did not meet the full definition: Nicaragua (2018), Vietnam (2022), and Sri Lanka (2023).
- Newly identified episodes (relative to Laeven and Valencia 2020)
  - Angola — 2015
  - Azerbaijan — 2015
  - Chad — 2015
  - Equatorial Guinea — 2015
  - Ghana — 2015
  - Kazakhstan — 2014
  - Lebanon — 2019
  - Nicaragua* — 2018
  - Republic of the Congo — 2018
  - São Tomé and Príncipe — 2016
  - Sri Lanka* — 2023
  - Tajikistan — 2016
  - Vietnam* — 2022
  - (* Denotes borderline cases)
- Character of new episodes
  - Predominantly triggered by significant negative shocks outside the financial sector (e.g., sharp decline in oil prices starting late 2014 affecting oil-producing countries such as Angola, Chad, Azerbaijan).
  - Contrast with global financial crisis episodes where distress originated within the financial sector.
- Distributional facts
  - Most countries experienced at least one systemic banking crisis during 1970–2025; few countries experienced more than two episodes (including borderline cases): Argentina (4), Chad (3), Democratic Republic of Congo (3), Nicaragua (3), Ukraine (3).
- 2023 special events (March 2023 in US and Switzerland)
  - United States:
    - Failures in rapid succession: Silicon Valley Bank (SVB), Signature Bank, First Republic.
    - Combined deposits of the three banks: US$ 367 billion, or 2.0 percent of total US bank deposits.
    - SVB and Signature Bank had about 90% uninsured deposits; bank runs occurred at unprecedented speed.
    - Federal Reserve established a new liquidity facility; systemic risk exception invoked; uninsured depositors in SVB and Signature Bank were fully protected by the FDIC, backed by a US Treasury guarantee.
    - First Republic failed on May 1, 2023 (closed after downgrades and large-scale deposit withdrawals); FDIC sold the bank and almost all its assets to J.P. Morgan Chase.
    - Assessment: failures did not meet the paper’s criteria for a systemic banking crisis:
      - Peak liquidity support amounted to about 1% of total US banks liabilities (below 5 percent threshold).
      - No fiscal costs from bank restructuring.
      - Government guarantees applied explicitly to affected institutions but not to the banking system.
  - Switzerland (Credit Suisse):
    - Credit Suisse represented 14.2 percent of Swiss total banking sector assets.
    - Swiss National Bank injected 168 billion Swiss francs (around US$ 185 billion) in emergency liquidity support (around 20% of Swiss GDP).
    - Taken over by UBS with a 9 billion Swiss franc state guarantee on the value of the bank’s assets.
    - Liquidity support provided by SNB represented 4.5 percent of total banking system liabilities in Switzerland (below 5 percent threshold).
    - Aside from state guarantee on bank’s liabilities to support takeover, there were no fiscal outlays from bank restructuring costs.
    - Assessment: despite being a systemic institution, the event does not qualify as a systemic banking crisis under the paper’s definition due to speedy resolution and thresholds not met.

### V. Policy Interventions during Crises
- Overview of instruments used
  - Both high-income and low and middle-income countries heavily rely on liquidity support during crises.
  - High-income countries tend to deploy a wider array of instruments (e.g., coordinated central bank swap facilities) and make greater use of significant guarantees on bank liabilities.
  - Deposit freezes and bank holidays are rare but recent cases include deposit freezes: Cyprus (2013), Ukraine (2014), Greece (2015), Lebanon (2019); bank holidays: Cyprus, Greece, Lebanon, Sri Lanka.
- Stylized facts
  - Peak liquidity support:
    - Liquidity support measured as ratio of central bank claims on financial sector to deposits and foreign liabilities.
    - Median peak liquidity support ratio in the sample reaches 20 percent.
  - Guarantees:
    - Significant guarantees on bank liabilities are more common among high-income countries.
    - Explanations: stronger institutions, greater fiscal space, lower prevalence of foreign currency–denominated liabilities.
- Recapitalization and restructuring
  - Private or public recapitalization, resolution of insolvent institutions, and nationalization are common.
  - Bank recapitalization with public funds is often the most important component of direct fiscal costs.
  - Government capital injections often involve preferred and common equity and conditions (board seats, dividend restrictions); governments can retain ownership for many years.
- Regulatory context and crisis toolkit
  - Post-GFC reforms (Basel III) enlarged the crisis-response toolkit via releasable capital buffers (capital conservation buffer, countercyclical capital buffer - CCyB) and enhanced liquidity standards (Liquidity Coverage Ratio - LCR).
  - These frameworks allow authorities to release buffers in downturns, supporting continued lending without breaching regulatory thresholds.
  - Value of these reforms was evident during the COVID-19 crisis in avoiding a wave of systemic banking crises.

### VI. Crises Outcomes
- Metrics reported (detailed data in Appendix I Table A2)
  - Direct fiscal costs: fiscal outlays linked to government interventions in the banking system, normalized by nominal GDP of the year of outlay.
  - Peak nonperforming loans (NPLs): peak NPL ratio in percent of total loans.
  - Crisis duration: number of years between start and end of crisis (end defined as the year before both real GDP growth and real credit growth are positive for at least two consecutive years); durations truncated at five years.
  - Output losses: 4-year cumulative deviation of output from pre-crisis trends (cumulative over [T, T+3], expressed as percent of one year’s trend GDP).
- A. Fiscal Costs
  - Measurement details:
    - Fiscal costs include capital injections, operating costs of asset management companies, exercised public guarantees that led to outlays, and other direct fiscal rescue costs.
    - Excludes government guarantees as they are not ex-post outlays; excludes emergency liquidity assistance.
    - Fiscal costs are reported in percent of GDP; nominal outlays converted to domestic currency and divided by nominal GDP of corresponding year.
  - Summary statistics and distributions:
    - Median gross fiscal cost for high-income country crises: 6.7 percent of GDP (med = 6.67).
    - Median gross fiscal cost for low and middle-income country crises: 7.5 percent of GDP (med = 7.50).
    - Histograms show substantial variation across episodes.
  - Data sources: official country publications, supranational agencies, IMF staff reports; fiscal costs for earlier episodes taken from Laeven and Valencia (2020) except newly identified episodes in this paper.
  - Notable data point: For Lebanon (2019), direct fiscal costs are not available; 2023 Article IV consultation staff report noted estimated overall losses in the financial system at about US$70 bn (more than 300 percent of the estimated 2022 GDP).
- B. Peak Nonperforming Loans (NPLs)
  - Caveats: NPL measures affected by regulatory differences, forbearance, and definition heterogeneity.
  - Median peak NPLs:
    - High-income countries: med = 11.14 (peak NPL in percent of total loans).
    - Low and middle-income countries: med = 30.00 (peak NPL in percent of total loans).
- C. Systemic Banking Crises Duration
  - End-date rule: year before both real GDP growth and real credit growth are positive for at least two consecutive years; credit measured as bank credit to private sector (national currency) from IFS deflated by CPI from WEO; GDP from WEO; when credit data unavailable, end date determined by GDP growth rule.
  - Truncation: crisis durations truncated at five years to avoid upward bias.
  - Distribution:
    - About two-thirds of crises ended in less than five years.
    - Crises in high-income countries tend to be more persistent, often lasting five years or more.
    - Most crises in low and middle-income countries lasted four years or less.
  - Explanations for duration differences: crisis severity (global financial crisis episodes concentrated in high-income countries), complexity of larger financial systems, use of monetary and fiscal policy which may reduce urgency of active restructuring.
- D. Output Losses
  - Measurement:
    - Output losses computed as cumulative sum of differences between actual and trend real GDP over [T, T+3], expressed as percent of trend real GDP.
    - Trend computed via HP filter (λ=100 for annual data) on log real GDP over [T-20, T-1] or longest available series with at least 4 pre-crisis observations; real GDP extrapolated using trend growth rate.
    - For crises starting in 2013 or later, calculations use real GDP data from the fall 2025 WEO; for earlier crises, output losses taken from Laeven and Valencia (2020).
  - Distribution:
    - Output losses in high-income countries tend to be much larger than in low and middle-income countries.
    - Median cumulative output loss:
      - High-income countries: med = 31.75 (cumulative output loss in percent of GDP).
      - Low and middle-income countries: med = 10.80 (cumulative output loss in percent of GDP).
  - Caveat: output losses reflect losses around crisis episodes but are not solely caused by banking crises; may include multiple concurrent shocks.

### VII. Conclusions
- Methodology and contribution
  - Textual analysis tools applied to IMF staff reports and supplementary sources improved detection of potential crisis episodes; expert evaluation and quantitative thresholds determine final inclusion and dating.
  - The update extends the Laeven and Valencia database through end-2025, enhancing comprehensiveness across income levels and documenting policy responses.
- Key findings
  - Systemic banking crises remained relatively rare between 2020 and 2025 despite major global shocks, reflecting increased resilience of the global banking sector and the impact of post-GFC reforms and policy support.
  - Banking turmoil in 2023 (US and Switzerland) underscores the continuing relevance of research on crisis avoidance, mitigation, and resolution even when events do not meet defined systemic thresholds.
- Uses and outlook
  - The database supports policy and analytical work: understanding cross-country vulnerability patterns, evaluating resolution frameworks, assessing macroeconomic effects of crises, and evaluating effectiveness of containment measures.

*Source: Authors’ calculations.*

### References

### References

### Journal articles, working papers, and books
- Abbassi, Puriya, Rajkamal Iyer, José-Luis Peydró, and Francesc R. Tous. 2016. Securities Trading by Banks and Credit Supply: Micro-Evidence from the Crisis. Journal of Financial Economics, 121(3), 569–594.
- Alfaro, Laura, Manuel García-Santana, and Enrique Moral-Benito. 2017. On the Direct and Indirect Real Effects of Credit Supply Shocks. NBER Working Paper No. 23450, National Bureau of Economic Research.
- Ashcraft, Adam B. 2005. Are Banks Really Special? New Evidence from the FDIC-Induced Failure of Healthy Banks. American Economic Review, 95(5), 1712–1730.
- Baron, Matthew, Emil Verner, and Wei Xiong. 2021. Banking Crises without Panics. The Quarterly Journal of Economics, 136(1), 51–113.
- Bernanke, Ben S., and Mark Gertler. 1987. Financial Fragility and Economic Performance. NBER Working Paper No. 2318. National Bureau of Economic Research. https://www.nber.org/system/files/working_papers/w2318/w2318.pdf
- Dell’Ariccia, Giovanni, Enrica Detragiache, and Raghuram Rajan. 2008. The Real Effect of Banking Crises. Journal of Financial Intermediation, 17(1), 89–112.
- Giannetti, Mariassunta, and Andrei Simonov. 2013. On the Real Effects of Bank Recapitalizations: Self-Fulfilling Prophecies and the Credit Crunch. Journal of Financial Economics, 108(1), 150–170.
- Homar, Timotej, and Sweder van Wijnbergen. 2017. On Bank Recapitalization and Stagnation 20 Years After the Japanese Crisis. Journal of Financial Stability, 30, 1–17.
- Kroszner, Randall S., Luc Laeven, and Daniela Klingebiel. 2007. Banking Crises, Financial Dependence, and Growth. Journal of Financial Economics, 84(1), 187–228.
- Laeven, Luc and Fabian Valencia. 2020. Systemic Banking Crises database II. IMF Economic Review, 68(2), 307–361.
- Laeven, Luc, and Fabian Valencia. 2013. Systemic Banking Crises Database. IMF Economic Review, 61(2), 225–270.
- Laeven, Luc, and Fabian Valencia. 2012. Systemic Banking Crises Database: An Update. IMF Working Paper No. 12/163. Washington, DC: International Monetary Fund. https://doi.org/10.5089/9781475505054.001
- Laeven, Luc, and Fabian Valencia. 2010. Resolution of Banking Crises: The Good, the Bad, and the Ugly. IMF Working Paper 10/146. Washington, DC: International Monetary Fund.
- Laeven, Luc, and Fabian Valencia. 2008. Systemic Banking Crises: A New Database. IMF Working Paper No. 08/224. Washington, DC: International Monetary Fund. https://www.imf.org/en/publications/wp/issues/2016/12/31/systemic-banking-crises-a-new-database-22345
- Peek, Joe, and Eric Rosengren. 1997. The International Transmission of Financial Shocks: The Case of Japan. American Economic Review, 87(4), 495–505.
- Philippon, Thomas, and Philipp Schnabl. 2013. Efficient Recapitalization. The Journal of Finance, 68(1), 1–42.
- Reinhart, Carmen and Kenneth Rogoff. 2009. This Time Is Different: Eight Centuries of Financial Folly. Princeton University Press. https://psycnet.apa.org/record/2009-18858-000
- Romer, Cristina and David Romer. 2017. New Evidence on the Aftermath of Financial Crises in Advanced Countries. American Economic Review, 107(10), 3072–3118. https://doi.org/10.1257/aer.20150320
- Sandri, Damiano, and Fabian Valencia. 2013. Financial Crises and Recapitalizations. IMF Working Paper No. 13/59. Washington, DC: International Monetary Fund.
- Schularick, Moritz, and Alan Taylor. 2012. Credit Booms Gone Bust: Monetary Policy, Leverage Cycles, and Financial Crises, 1870–2008. American Economic Review, 102(2), 1029–1061.
- Stone, Mark, Kotaro Ishi, and Erik Nordström. 2011. The Role of Central Banks in Financial Stability: How Has It Changed? IMF Working Paper 11/26. Washington, DC: International Monetary Fund.
- Valencia, Fabian. 2014. Monetary Policy, Bank Leverage, and Financial Stability. Journal of Economic Dynamics and Control, 47, 20–38.
- Van Den Heuvel, Skander J. 2006. The Welfare Cost of Bank Capital Requirements. Journal of Monetary Economics, 55(2), 298–320.

### IMF reports, staff notes, and working papers
- Claessens, Stijn, Giovanni Dell’Ariccia, Deniz Igan, and Luc Laeven. 2011. Lessons and Policy Implications from the Global Financial Crisis. IMF Working Paper 11/44. Washington, DC: International Monetary Fund.
- Claessens, S., Kose, M.A., Laeven, L., and Valencia, F. (eds.) (2014). Financial Crises: Causes, Consequences, and Policy Responses. Washington, DC: International Monetary Fund.
- Igan, Deniz O., Hala Moussawi, Alexander F. Tieman, Aleksandra Zdzienicka, Giovanni Dell’Ariccia, and Paolo Mauro. 2019. The Long Shadow of the Global Financial Crisis: Public Interventions in the Financial Sector. IMF Working Paper 19/164. Washington, DC: International Monetary Fund.
- International Monetary Fund. 2017. Spain: Financial System Stability Assessment (IMF Country Report No. 17/321). Washington, DC: International Monetary Fund. https://www.imf.org/en/publications/cr/issues/2017/10/06/spain-financial-system-stability-assessment-45321
- International Monetary Fund. 2020. Global Financial Stability Report: Bridge to Recovery. Washington, DC: International Monetary Fund. https://www.imf.org/en/publications/gfsr/issues/2020/10/23/global-financial-stability-report-october-2020-bridge-to-recovery-49686
- International Monetary Fund. 2021a. Global Financial Stability Report: Preempting a Legacy of Vulnerabilities. Washington, DC: International Monetary Fund. https://www.imf.org/en/publications/gfsr/issues/2021/04/06/global-financial-stability-report-april-2021
- International Monetary Fund. 2021b. Global Financial Stability Report: COVID-19, Crypto, and Climate: Navigating Challenging Transitions. Washington, DC: International Monetary Fund. https://www.imf.org/en/publications/gfsr/issues/2021/10/12/global-financial-stability-report-october-2021
- International Monetary Fund. 2022a. Global Financial Stability Report: Shockwaves from the War in Ukraine Test the Financial System’s Resilience. Washington, DC: International Monetary Fund. https://www.imf.org/en/publications/gfsr/issues/2022/04/18/global-financial-stability-report-april-2022-516158
- International Monetary Fund. 2022b. Global Financial Stability Report: Navigating the High-Inflation Environment. Washington, DC: International Monetary Fund. https://www.imf.org/en/publications/gfsr/issues/2022/10/04/global-financial-stability-report-october-2022-navigating-the-high-inflation-environment-523390
- Igan, Deniz O., Hala Moussawi, Alexander F. Tieman, Aleksandra Zdzienicka, Giovanni Dell’Ariccia, and Paolo Mauro. 2019. The Long Shadow of the Global Financial Crisis: Public Interventions in the Financial Sector. IMF Working Paper 19/164. Washington, DC: International Monetary Fund.

### Banking supervision, regulatory, and central bank reports
- Abboud, Alice, Elizabeth Duncan, Akos Horvath, Diana Iercosan, Bert Loudis, Francis Martinez, Timothy Mooney, Ben Ranish, Ke Wang, Missaka Warusawitharana, and Carlo Wix. 2021. “COVID-19 as a Stress Test: Assessing the Bank Regulatory Framework,” Finance and Economics Discussion Series 2021-024. Washington: Board of Governors of the Federal Reserve System.
- Basel Committee on Banking Supervision. 2021. Early Lessons from the COVID-19 Pandemic on the Basel Reforms. Bank for International Settlements. https://www.bis.org/bcbs/publ/d521.pdf
- Basel Committee on Banking Supervision. 2023. Report on the 2023 Banking Turmoil. Bank for International Settlements. https://www.bis.org/bcbs/publ/d555.pdf
- Bank for International Settlements. (2021). Annual Economic Report 2021. https://www.bis.org/publ/arpdf/ar2021e.pdf
- Board of Governors of the Federal Reserve System. 2023. Review of the Federal Reserve’s Supervision and Regulation of Silicon Valley Bank. https://www.federalreserve.gov/publications/files/svb-review-20230428.pdf
- Swiss Financial Market Supervisory Authority (FINMA). 2023. Report on the Credit Suisse Crisis. https://www.finma.ch/en/news/2023/12/20231219-mm-cs-bericht/

### Crisis management, databases, and broader assessments
- Acharya, Viral, Matthew Richardson, Kermit Schoenholtz, and Bruce Tuckman (Eds.). 2023. SVB and Beyond: The Banking Stress of 2023. CEPR Press. https://cepr.org/publications/books-and-reports/svb-and-beyond-banking-stress-2023
- Angeloni, Ignacio, Stijn Claessens, Amit Seru, Sascha Steffen, and Beatrice Weder di Mauro. 2024. Much Money, Little Capital, and Few Reforms: The 2023 Banking Turmoil. Geneva Reports on the World Economy 27, International Center for Monetary and Banking Studies (ICMB) and Centre for Economic Policy Research (CEPR).
- Pazarbasioglu, C., Laeven, L., Croitoru, O., Claessens, S., Valencia, F., Dobler, M., and Seal, K. 2011. Crisis Management and Resolution: Early Lessons from the Financial Crisis. IMF Staff Discussion Note 11/05

*IMF WORKING PAPERS Systemic Banking Crises Database: 1970-2025 — References section*

### Appendix I. Banking Crises Episodes and Policy

### Appendix I. Banking Crises Episodes and Policy Responses

### Banking crises episodes: scope and key metrics (Table A1)
- Coverage: Systemic banking crises database spanning 1970–2025 with country-level crises episodes listed by Start and End year.
- Metrics reported for each episode:
  - Output loss: "cumulative sum of the differences between actual and trend real GDP over the period [T, T+3], expressed in percent of trend real GDP" (trend via HP filter λ=100 over [T-20, T-1]).
  - Fiscal costs: "outlays directly related to the restructuring of the financial sector" (percent of GDP).
  - Peak liquidity: "ratio of central bank claims (and liquidity provided by the treasury if any) on deposit money banks to total deposits and liabilities to non-residents."
  - Peak NPLs: "In percent of total loans."
  - Crisis duration truncated at 5 years where noted (footnote 5).
- Selected episode values (preserve original numeric precision exactly):
  - Albania 1994: Peak liquidity 7.6; Peak NPLs 26.8.
  - Algeria 1990–1994: Output loss 5/; Fiscal costs 41.4; Peak liquidity 37.6; Peak NPLs 30.0.
  - Argentina 2001–2003: Output loss 71.0; Fiscal costs 9.6; Peak liquidity 22.9; Peak NPLs 20.1.
  - Austria 2008–2012: Fiscal costs 19.2; Output loss 5/; Peak liquidity 10.0; Peak NPLs 4.1.
  - Brazil 1994–1998: Output loss 0.0; Fiscal costs 13.2; Peak liquidity 20.1; Peak NPLs 16.0.
  - Chile 1981–1985: Output loss 5/; Fiscal costs 8.6; Peak liquidity 61.2; Peak NPLs 35.6.
  - China 1998: Output loss 19.4; Fiscal costs 18.0; Peak liquidity 62.0; Peak NPLs 20.0.
  - Denmark 2008–2009: Output loss 35.0; Fiscal costs 5.9; Peak liquidity 17.7; Peak NPLs 5.95.
  - Greece 2008–2012: Output loss 64.9; Fiscal costs 28.7; Peak liquidity 61.7; Peak NPLs 37.1.
  - Iceland 2008–2012: Output loss 34.5; Fiscal costs 37.6; Peak liquidity 33.8; Peak NPLs 61.2.
  - Ireland 2008–2012: Output loss 107.7; Fiscal costs 37.6; Peak liquidity 18.1; Peak NPLs 25.7.
  - Korea 1997–1998: Output loss 57.6; Fiscal costs 31.2; Peak liquidity 27.4; Peak NPLs 17.7.
  - Latvia 2008–2012: Output loss 93.9; Fiscal costs 8.1; Peak liquidity 3.6; Peak NPLs 15.9.
  - Norway 1991–1993: Output loss 5.1; Fiscal costs 2.7; Peak liquidity 16.9; Peak NPLs 16.4.
  - Spain 2008–2012: Output loss 38.8; Fiscal costs 5.4; Peak liquidity 33.5; Peak NPLs 9.4.
  - Sweden 1991–1995: Output loss 32.9; Fiscal costs 3.6; Peak liquidity 3.1; Peak NPLs 13.0.
  - United Kingdom 2007–2011: Output loss 25.3; Fiscal costs 8.8; Peak liquidity 3.4; Peak NPLs 4.0.
  - United States 2007–2011: Output loss 30.0; Fiscal costs 4.5; Peak liquidity 4.7; Peak NPLs 5.0.
  - Uruguay 2002–2005: Output loss 66.1; Fiscal costs 20.0; Peak liquidity 12.8; Peak NPLs 36.3.
  - Vietnam 1997: Output loss 0.0; Fiscal costs 10.0; Peak liquidity 64.9; Peak NPLs 35.0.
- Footnotes and special codes:
  - Footnote 5: durations of crises truncated at 5 years, starting with the first crisis year.
  - Footnote 6: borderline cases flagged in the dataset.
  - Source list: WEO, IFS, IMF Staff reports, IMF Financial Soundness Indicators, Laeven and Valencia (2020), and authors’ calculation.

### Banking crisis policy responses (Table A2)
- Table A2 documents policy tools used during crises for a subset of episodes, with columns including:
  - Deposit Freeze: Date and Duration (months).
  - Bank Holiday: Date and Duration (days).
  - Guarantees on Bank Liabilities: Start, Duration (months), and Coverage.
  - Bank Nationalization (Y/N).
  - Asset Purchase (In percent of GDP).
  - Bank Recapitalization (In percent of GDP).
  - IMF Program (year where applicable).
- Types of policy responses recorded (with exact example entries as reported):
  - Deposit freezes and bank holidays:
    - Argentina Dec-89: Deposit Freeze date 28-Dec-1989 duration 120 (days); Bank Holiday date 1-Jan-1990 duration 4 (days).
    - Cyprus Jun-11: Deposit Freeze date 28-Mar-2013 duration 14 (months); Bank Holiday date 18-Mar-2013 duration 8 (days).
    - Iceland Sep-08: Deposit Freeze (no date shown here) with unlimited coverage to domestic deposits; Asset Purchase 24.30 (percent of GDP); IMF Program 2008.
  - Guarantee programs (start, duration, coverage specifics):
    - Austria Sep-08 / Dec-08: "Unlimited coverage to depositors, bank and non-bank bonds."; Bank Recapitalization 3.96 (percent of GDP).
    - Azerbaijan Sep-15 / Mar-16 duration 57 (months): "All depositors’ funds irrespective of the amount and currency denomination"; Bank Recapitalization 1.00 (percent of GDP).
    - Belgium Sep-08 / Oct-08: "Deposit-like insurance instruments. Interbank loans" ; Bank Recapitalization 5.80 (percent of GDP).
    - Denmark Sep-08 / Feb-09: "Deposits and unsecured claims of PCA banks."; Bank Recapitalization 2.80 (percent of GDP).
    - Greece Sep-08 / Oct-08: Oct-08 start; Bank Recapitalization 25.40 (percent of GDP); IMF Programs 2010, 2012.
    - Indonesia Nov-97 / Jan-98 duration 78 (months): "All liabilities of domestic banks (excluding shareholders’ capital, subordinated debt, and related-parties deposits)."; Asset Purchase Y; Bank Recapitalization 37.30 (percent of GDP); IMF Program 1998.
    - Ireland Sep-08 / Sep-08: "Unlimited coverage to most liabilities of 10 banks."; Asset Purchase Y; Bank Recapitalization 37.13 (percent of GDP); IMF Program 2010.
    - Japan Nov-97 / Nov-97 duration 89 (days): "All deposits, including interbank deposits"; Asset Purchase Y; Bank Recapitalization 6.61 (percent of GDP).
    - Malaysia Jul-97 / Jan-98 duration 91 (days): "Deposits only of commercial banks, finance companies and merchant banks, including overseas branches of domestic banking institutions."; Asset Purchase Y; Bank Recapitalization 16.40 (percent of GDP).
    - Mexico Dec-94 / Dec-93 duration 109 (days): "All bank liabilities except subordinated debt"; Asset Purchase Y; Bank Recapitalization 3.80 (percent of GDP); IMF Program 1995.
    - Norway Oct-91: Deposit Freeze Y; Bank Recapitalization 2.61 (percent of GDP).
    - Russia Sep-08 / Nov-08: "Interbank borrowing for qualifying banks."; Bank Recapitalization 2.30 (percent of GDP).
    - Sweden Sep-91 / Sep-92 duration 46 (months): "All liabilities, except for shareholders"; Asset Purchase Y; Bank Recapitalization 1.85 (percent of GDP).
    - Thailand Jul-97 / Aug-97 duration 89 (days): "Deposits, contingent and foreign liabilities (excluding shareholders’ capital and subordinated debt) of banks and finance companies."; Asset Purchase Y; Bank Recapitalization 18.80 (percent of GDP); IMF Program 1998.
    - Turkey Nov-00 / Dec-00 duration 43 (days): "All liabilities (including contingent) of domestically incorporated banks except for owners' deposits, deposits linked to criminal activities, subordinated debt, and equity"; Asset Purchase Y; Bank Recapitalization 24.50 (percent of GDP); IMF Program 2000.
    - United Kingdom Sep-07 / Oct-08: "Guarantee on short-to-medium term debt; blanket guarantee on Northern Rock and Bradford & Bingley wholesale deposits."; Asset Purchase Y; Bank Recapitalization 5.00 (percent of GDP).
    - United States Dec-07 / Oct-08: "Money market funds (capped at US$50 billion); full guarantee on transaction deposits; newly issued senior unsecured debt."; Asset Purchase Y; Bank Recapitalization 3.60 (percent of GDP).
  - Nationalizations:
    - Examples where Bank Nationalization = Y: Angola 2015 entry indicates "Y" for nationalization with IMF Program 2018; Cyprus Jun-11 nationalization Y with Asset Purchase and Bank Recapitalization 18.90 (percent of GDP) and IMF Program 2013.
  - Asset purchases and recapitalizations (reported in percent of GDP):
    - Cyprus Jun-11: Asset Purchase Y; Bank Recapitalization 18.90 (percent of GDP); IMF Program 2013.
    - Finland Sep-91 / Feb-93: Asset Purchase Y; Bank Recapitalization 8.63 (percent of GDP).
    - Iceland Sep-08: Asset Purchase Y; Bank Recapitalization 24.30 (percent of GDP); IMF Program 2008.
    - Ireland Sep-08: Asset Purchase Y; Bank Recapitalization 37.13 (percent of GDP); IMF Program 2010.
    - Japan Nov-97: Asset Purchase Y; Bank Recapitalization 6.61 (percent of GDP).
    - Korea Aug-97: Asset Purchase Y; Bank Recapitalization 19.31 (percent of GDP); IMF Program 1998.
    - Turkey Nov-00: Asset Purchase Y; Bank Recapitalization 24.50 (percent of GDP); IMF Program 2000.
- Notes and data caveats included in Table A2:
  - Table A2 includes only a subset of the crises from Table A1, limited by availability of detailed policy response information.
  - Footnote 2: Saigon Commercial Bank (SCB) was placed under special control measures under the central bank in 2023 without transfer of ownership (specific to Vietnam 2022/2023 entry).
  - Data sources: Laeven and Valencia (2020), and authors' calculations.

*Source: Appendix I. Banking Crises Episodes and Policy Responses, IMF Working Papers — Systemic Banking Crises Database: 1970-2025.*

### Appendix II.  Implementation of Textual Analysis

### Appendix II.  Implementation of Textual Analysis

### Text preprocessing and normalization
- Raw text cleaning and filtering steps:
  - Splitting the raw text into sentences.
  - Removing punctuation and unwanted characters.
  - Eliminating elements such as section headers, titles, table rows, hyperlinks, and footnotes.
- Sentence-level processing:
  - Each sentence was tokenized and lemmatized.
  - Converting sentences to lowercase.
  - Removing stopwords.
  - Retaining only valid English words.
- Lexical normalization tools and implementation:
  - Used the WordNet lemmatizer from the Natural Language Toolkit (NLTK) in Python for lexical normalization.
  - All steps are programmed in Python.
- Output storage:
  - The lemmatized words were stored in the dataset.

### Dictionary for Textual Analysis (selected entries)
- Core banking and crisis terms included (examples shown verbatim):
  - administrative, liquidation, asset backed security, asset fall, asset guarantee, asset purchase
  - assist bank funding, assisted merger, authority responded, automatic stabilizer, bad bank, bailout
  - bank, bank asset purchase, bank assistance, bank authority, bank bailout, bank capitalization
  - bank closure, bank collapse, bank collateral, bank consolidation, bank crisis, bank default
  - bank distress, bank emergency, bank exposure, bank failure, bank freeze, bank funding assistance
  - bank guarantee, bank holiday, banking collapse, banking crisis, banking distress, banking regulation
  - banking sector vulnerability, banking support, banking system, banking turmoil, bank involvement
  - bank liquidation, bank loss, bank management, bank measure, bank nationalization, bank obligation
  - bank panic, bank problem, bank recapitalization, bank reform, bank rehabilitation, bank rescue
  - bank resolution, bank restructuring, bank run, bank squeeze, bank stimulus, bank suffered
  - bank support, bank turmoil, bank vulnerability, bank writedown, biggest bank, bridge bank
  - blanket deposit guarantee, blanket guarantee, capital adequacy, capital buffer, capital crunch
  - capital erosion, capital failure, capital infusion, capital injection, capital loss, capital market
  - capital reserve, capital requirement, capital risk, capital shortage, capital support, compensation scheme
  - confidence dropped, contagion, contagion risk, contingency plan, contingency planning
  - central bank intervened, central bank intervention, clearing bank, credit collapse, credit crunch
  - credit default, credit failure, credit guarantee, credit instability, credit loss, credit risk
  - credit squeeze, crisis management, crisis response, debt amortization, deep recession, default
  - deposit guarantee, deposit freeze, deposit insurance, depositor concern, depositor reimbursement
  - depositor protection, depositor withdrawal, deposit protection, deposit reimbursement, deposit withdrawal
  - deteriorating asset quality, deteriorating economic outlook, deterioration, downsize, downsizing
  - ecb lending, economic challenge, economic difficulty, economic downturn, economic instability
  - economic sector fragility, economic slowdown, economic weakness, emergency intervention
  - emergency liquidity, emergency measure, emergency support, equity injection, financial collapse
  - financial contagion, financial crisis, financial default, financial distress, financial loss
  - financial market disruption, financial meltdown, financial problem, financial risk, financial sector guarantee
  - financial stabilization, financial stabilization package, financial stress, financial system risk
  - financial trouble, financial turmoil, financial volatility, financial vulnerability, funding guarantee
  - funding pressure, funding risk, funding tightness, government bailout, government freeze
  - government funding assistance, government guarantee, government intervened, government intervention
  - government involvement, government pledged, government measure, government nationalization
  - government recapitalization, government rescue, government support, guarantee, high leverage ratio
  - idiosyncratic risk, illiquidity, impaired loan, increased risk, increased uncertainty, inject capital
  - injected capital, inject fund, injection capital, injecting fund, insolvency, insolvent, insurance company
  - interbank exposure, interbank funding, interbank lending, interbank loan, interbank market
  - intervene, intervention, intervention framework, intervention measure, involvement bank, large bank
  - large financial institution, large private bank, large public bank, largest bank, last resort, liquidate
  - liquidity assistance, liquidity monitoring, liquidity pressure, liquidity risk, liquidity shortage
  - liquidity support, loan impairment, macroprudential policy, major bank, major financial institution
  - management change, market authority, market distortion, market resolution, market supervision
  - market support, market turbulence, market turmoil, national bank, national central bank, nationalized
  - non performing, npl, npls, nonperforming loan, panic, policy intervention, policy measure
  - public guarantee, public sector support, purchase assets, recapitalization, recapitalize bank, recovery
  - reduced credibility, regulatory forbearance, regulatory guideline, regulatory intervention
  - regulatory measure, regulatory requirement, regulatory response, reserve, rescue, resolution measure
  - restructure, restructuring, rising vulnerability, risk financial system, rollover risk, safety net
  - securitization, securities regulator, shareholder dilution, solvency concern, state guarantee
  - state support, stimulus, strengthen financial stability, stress test, stress testing, structural weakness
  - support market liquidity, support measure, support package, systematic intervention, systemic concern
  - systemic disruption, systemic pressure, systemic risk, tail risk, tightening lending, toxic asset
  - troubled bank, troubled financial institution, troubled institution, undercapitalisation, undercapitalization
  - unsecured financing, unsound bank, vulnerability banking sector, vulnerability financial system, weak bank
  - wholesale funding

### Contextual metadata appearing in the source
- Title/Database: Systemic Banking Crises Database: 1970-2025
- Working Paper number: Working Paper No. WP/2026/094

*Source: Appendix II.  Implementation of Textual Analysis (from the provided PDF content).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2026/english/wpiea2026094-source-pdf.pdf_