## _wp08226

## Source details

**Canonical URL:** [_wp08226](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp08226.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp08226.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp08226.pdf.json)

---

### I. Introduction — purpose and approach
- Proposes a "thresholds method" to identify credit booms by splitting real credit per capita into cyclical and trend components and labeling episodes where credit exceeds its long-run trend by more than a specified multiple of the country-specific cyclical standard deviation.
- Key features of the method:
  - Boom and duration thresholds are proportional to each country’s standard deviation of credit over the business cycle, so booms are country-specific “unusually large” cyclical credit expansions.
  - Baseline boom threshold φ = 1.75; sensitivity checks for φ = 1.5 and φ = 2.
  - Starting and ending thresholds (φs and φe) baseline set to 1; alternative values tried include 0, ¼, ½ and ¾.
- Sample and scope:
  - Applied to data for 48 (later described as 49) countries over the 1960-2006 period; split into industrial and emerging economies.
  - Constructs seven-year event windows around boom peaks to examine macro aggregates and firm- and bank-level financial indicators.

### II. Methodology — formal definition and implementation
- Formal definition (notation preserved):
  - lit denotes deviation from long-run trend in log real credit per capita for country i at date t.
  - σi(l) denotes the standard deviation of this cyclical component.
  - A credit boom is a contiguous period where lit ≥ φ·(σi(l)).
  - Peak date t̂ is where lit − φ·(σi(l)) is maximal within contiguous dates satisfying the boom condition.
  - Start date ts minimizes |lit − φs·(σi(l))| for ts < t̂; end date te minimizes |lit − φe·(σi(l))| for te > t̂.
- Trend and detrending:
  - Trend computed with Hodrick-Prescott (HP) filter with smoothing parameter set at 100 (standard for annual data). Sensitivity to other smoothing parameters examined.
- Credit measure and construction:
  - Credit measure is the sum of claims on the private sector by deposit money banks (IFS line 22d) plus, when available for the entire sample period for a country, claims by other financial institutions (IFS line 42d).
  - Real credit per capita calculated as the average of two contiguous end-of-year observations of nominal credit per capita deflated by their respective end-of-year CPI.

### III. Counts, magnitudes, and durations of credit booms
- Counts and sample caveats:
  - Identified 49 credit booms: 27 in industrial countries and 22 in emerging economies.
  - Four emerging economies (India, South Africa, Turkey, and Venezuela) had booms identified in 2006 but were excluded from event analysis because they were not completed.
- Magnitude and relative size:
  - At boom peaks, average expansion in real credit per capita reached almost 30 percent above trend in emerging economies, twice what is observed in industrial countries.
  - In multiples of country-specific standard deviations, booms in industrial and emerging economies are similar (emerging economies exhibit larger cyclical variability).
- Duration and phases:
  - Medians indicate EMs and ICs show booms with similar duration of about 6-7 years and upswings that last longer than downswings.
  - Using means, EMs appear to show longer and more asymmetric booms (driven largely by Sudden Stops cases).

### IV. Macro dynamics in seven-year event windows (empirical regularities)
- Variables analyzed (detrended with HP λ=100 unless noted): output (Y), private consumption (C), public consumption (G), investment (I), nontradables output (YN), real exchange rate (RER), current account-output ratio (CAY), capital inflows share of output (KI).
- Emerging economies (typical patterns and magnitudes, relative to trend):
  - Y, C and G: rise 2 to 4 percentage points above trend in build-up; fall 3 to 4 percent below trend in downswing.
  - Investment I: rises up to 18 percent above trend at t = -1 and drops below trend by a similar amount by t = 2.
  - Nontradables YN: rises to about 6½ percent above trend by t = 0 and drops to almost 3 percent below trend by t = 3.
  - Median RER: appreciates 9 percent above trend at t = -1, drops to about 4 percent below trend when boom unwinds.
  - CAY: declines to a deficit of about 2½ percentage points of GDP in expansion and rises to a surplus of 1½ percentage points of GDP in decline.
  - KI: rises by up to 3½ percentage points of GDP by t = -1 and then drops by 2¼ percentage points of GDP by t = 3.
- Industrial countries:
  - Similar cyclical pattern but smaller amplitudes.
  - Government consumption shows slight decline in expansion and slight rise in contraction (opposite of EMs).
  - Nontradables sector fluctuations much smaller; RER slightly depreciates in expansions rather than appreciations seen in EMs.
- Statistical significance:
  - Cross-section regressions show majority of mean and median estimates for Y, YN, C, and I are statistically significant at the 1 percent level for both ICs and EMs; G, RER and CA/Y often have larger standard errors.
- Comparison with output booms:
  - Conditioning on output booms yields little discernible credit cycle in EMs; in ICs there is a credit cycle but much less pronounced than during credit booms.

### V. Prices, asset markets, and heterogeneity
- Prices and asset markets:
  - Industrial countries: negligible inflation changes; equity and housing prices rise in build-up and fall in decline.
  - Emerging economies: mean inflation spikes after boom peaks driven by outliers; median inflation rises only 2 percent above trend.
  - Equity and housing price dynamics similar to ICs but larger in magnitude in EMs.
- Regional heterogeneity:
  - Industrial-country variation: Nordic countries show larger fluctuations in credit and macro variables than G7; some macro variables peak earlier than credit in Nordics.
  - Emerging-economy variation: credit expansions and consumption fluctuations larger in Latin America; current account reversals larger in Asia.

### VI. Frequency analysis — crises, preconditions, and exchange-rate regimes
- Credit booms and crises (seven-year window frequencies):
  - Emerging economies:
    - Banking crises total: 0.55
    - Currency crises total: 0.68
    - Sudden Stops total: 0.32
  - Industrial countries:
    - Banking crises total: 0.15
    - Currency crises total: 0.54
    - Sudden Stops total: 0.04
  - All countries:
    - Banking crises total: 0.33
    - Currency crises total: 0.60
    - Sudden Stops total: 0.17
- Preconditions (frequency of occurrence before boom peaks):
  - Large Capital Inflows (A): Industrial Countries 0.27; Emerging Market Economies 0.50; All 0.52.
  - Significant Productivity Gains (B): Industrial Countries 0.40; Emerging Market Economies 0.17; All 0.33.
  - Large Financial Sector Changes (C): Industrial Countries 0.33; Emerging Market Economies 0.22; All 0.27.
  - Memo: (A) & (B) & (C) joint occurrence All = 0.06.
- Exchange rate regimes (Reinhart and Rogoff (2004) classification):
  - Fixed and managed: Industrial Countries 77.78; Emerging Market Economies 73.91; All 76.00.
  - Dirty Floating: Industrial Countries 7.41; Emerging Market Economies 17.39; All 12.00.
  - Floating: Industrial Countries 7.41; Emerging Market Economies 0.00; All 4.00.
  - Mixed: Industrial Countries 7.41; Emerging Market Economies 8.70; All 8.00.
- Early-warning probabilities (starting threshold crossings):
  - Starting threshold = 1 standard deviation:
    - Probability of a credit boom = 16 percent for emerging economies, 22 percent for industrial countries, and 18 percent for all countries combined.
  - Starting threshold = ½ standard deviation:
    - Probability of a credit boom = 9 percent for emerging economies, 14 percent for industrial countries, and 11 percent for all countries combined.

### VII. Microeconomic indicators — firm-level evidence
- Firm-level indicators constructed (definitions preserved):
  - Tobin’s Q measures: Q1 and Q2 definitions as given.
  - Effective interest rate (ER) = ratio of total debt service to total debt obligations.
  - Profitability (PR) = return on assets in percent.
  - Total leverage: LBV, LMV.
  - Short-term leverage (SLMV), Working capital leverage (LWK).
  - Rajan-Zingales index (RZ).
- Dynamics during credit booms:
  - Leverage:
    - All leverage ratios rise in the build-up phase and collapse in the declining phase.
    - Amplitude larger in emerging economies than in industrial countries.
    - Specific magnitudes:
      - LBV rises about 20 percentage points in emerging economies from its upswing minimum to peak (compared with 12 percentage points in industrial countries).
      - LWK rises almost 12 percentage points in emerging economies (compared with 2 percentage points in industrial countries).
  - Tobin’s Q and profitability:
    - Corporations start expansionary phase from high asset valuation and profitability, which decline at the boom peak and remain depressed afterward; magnitude larger in emerging economies.
  - Effective interest rate (ER):
    - Emerging markets: ER jumps about 500 basis points one year after the booms peak.
    - Industrial countries: ER is more stable and drops slightly after credit booms peak.
  - Rajan-Zingales (RZ) index:
    - Firms more dependent on external financing in the build-up; dependence falls sharply in downswing, with larger corrections in emerging economies.
- Regional and sectoral patterns:
  - Asia vs Latin America:
    - Firms in Asia are significantly more leveraged than in Latin America; leverage ratios fluctuate more sharply in Asia.
    - Tobin’s Q in Asia falls sharply in a continuous decline; Latin America shows a more modest decline.
    - Dependence on external financing falls sharply in both regions, larger correction in Latin America.
  - Tradables vs nontradables:
    - Leverage cycles larger for firms in the nontradables sector and synchronized with sharp real exchange rate cycles.
    - Tradables sector displays more stable access to credit than nontradables.

### VIII. Bank-level indicators and dynamics (emerging economies)
- Bank-level indicators constructed (definitions preserved):
  - ROA, NPL, LAR, CAR.
  - Basel benchmark: Minimum CAR recommended by the Basel Committee is 8 percent; supervisors in more vulnerable economies are encouraged to set higher minimum CARs (often in the 8-12 percent range).
- Dynamics during credit booms:
  - Lending and asset quality:
    - Lending activity high in expanding phase.
    - Non-performing loans rise from 2.5 percent a year before the peak to 10 percent two years after the peak.
  - Profitability:
    - ROA reaches highest level at 1.5 percent a year before the peak, followed by a collapse in the ending phase.
  - Capital adequacy:
    - CAR shows v-shaped dynamics: drops sharply late in the expanding phase and then rebounds early in the ending phase, ending at higher levels than when booms start.
    - CAR trough was just below 12 percent in the sample, indicating vulnerabilities in banking systems of these economies.

### IX. Macro-micro synchronization and stylized facts
- Synchronous pattern across booms:
  - Build-up phase: expansions in output, rising equity and housing prices, real currency appreciation, widening external deficits, increased firm leverage and external financing, rising bank lending and profitability.
  - Downswing: reversals in the above, increases in non-performing loans, declines in bank capital ratios and firm valuations.
- Differences between industrial and emerging economies:
  1. Fluctuations are larger, more persistent, and asymmetric in emerging economies; pattern particularly strong in the nontradables sector.
  2. Many recent emerging market crises were associated with credit booms, but not all booms end in crisis.
  3. Preconditions differ: EM booms often preceded by large capital inflows; IC booms more often follow TFP gains or financial reforms.

### X. Policy implications and recommendations
- Surveillance and early-warning:
  - Cross-country patterns indicate that monitored crossings of ½ or 1 standard deviation starting thresholds have measurable, though modest, predictive power (probabilities reported above) for subsequent credit booms and can be used as early-warning indicators.
- Policy signals to supplement credit measures:
  - Booms in output and expenditures.
  - Excessive real appreciation and/or expansion of the nontradables sector.
  - Large inflows of foreign capital (in EMs).
  - Fast TFP growth or domestic financial reforms (in ICs).
  - Increases in leverage and profitability ratios of corporations.
  - Weakening in the quality of banks’ assets.
- Corrective policy actions:
  - Use of corrective policy actions to prevent credit booms is important because the declining phase of credit booms is associated with recessions and a higher incidence of financial crises.
- Tailoring responses:
  - The strong association between managed/fixed exchange rate regimes and booms (about ¾ of booms) suggests exchange rate regime considerations are relevant.
  - Differing preconditions for booms in ICs (TFP gains, financial reforms) versus EMs (capital inflows) imply policy responses and preventive measures should be tailored to country context.

*Source: _wp08226 (IMF Working Paper content excerpts).*

### 1. Credit Booms: Duration ..............................................................................................

### 1. Credit Booms: Duration .......................................................................................................22

### Major sections (chapter headings)
- 1. Credit Booms: Duration .......................................................................................................22
- 2. Coincidence of Credit Booms with Output and Demand Booms ........................................23
- 3. Credit Booms: Statistical Significance of Event-Window Coefficients ..............................24
- 4. Credit Booms: Regional Features ........................................................................................26
- 5. Credit Booms and Crises .....................................................................................................27
- 6. Credit Booms: Potential Triggering Factors ........................................................................28
- 7. Credit Booms and Exchange Rate Regimes ........................................................................29

### Figures (titles)
- 1. Credit Booms: Seven-Year Event Windows........................................................................30
- 2. Relative Credit Booms .........................................................................................................31
- 3. Frequent of Credit Booms....................................................................................................32
- 4. Credit Booms in Chile: The Mendoza-Terrones Method ....................................................33
- 5. Credit Booms in Chile: The Gourinchas, Valdes, and Landeretche (CVL) Method ...........34
- 6. Credit Booms in Chile: Expanding vs. Conventional Trend ...............................................35
- 7. Credit Booms and Economic Activity .................................................................................36
- 8. Credit Booms and Domestic Demand..................................................................................37
- 9. Credit Booms and The Non-Tradables Sector .....................................................................38
- 10. Credit Booms, Current Account, and Capital Inflows .......................................................39
- 11. Credit Booms and Prices....................................................................................................40
- 12. Credit Booms: Corporate Leverage ...................................................................................41
- 13. Credit Booms: Tobin's Q, Interest Rate, and Profitability .................................................42
- 14. Credit Booms: Corporate External Financing....................................................................43
- 15. Credit Booms: Financial Indicators ...................................................................................44
- 16. Credit Booms: Corporate Leverage and External Financing in Emerging Market Countries............................................................................................45
- 17. Credit Booms: Bank Level Data Emerging Market Economies ........................................46

### Appendices and supplementary material
- Appendix I. Sample of Countries .............................................................................................................47
- Appendix II. Data Definitions and Sources ..............................................................................................48
- Appendix III. The Gourinchas-Valdes-Landerretche Thresholds Method...............................................49

*Source: _wp08226 - 1. Credit Booms: Duration .......................................................................................................22*

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

### _wp08226 - References..............................................................................................................

### I. Introduction — purpose and approach
- Proposes a "thresholds method" to identify credit booms by splitting real credit per capita into cyclical and trend components and labeling episodes where credit exceeds its long-run trend by more than a specified multiple of the country-specific cyclical standard deviation.
- Key features of method:
  - Boom and duration thresholds are proportional to each country’s standard deviation of credit over the business cycle, so booms are country-specific “unusually large” cyclical credit expansions.
  - Baseline boom threshold φ = 1.75; sensitivity checks for φ = 1.5 and φ = 2.
  - Starting and ending thresholds (φs and φe) baseline set to 1; alternative values tried include 0, ¼, ½ and ¾.
- Sample and scope:
  - Applied to data for 48 (later described as 49) countries over the 1960-2006 period; split into industrial and emerging economies.
  - Constructs seven-year event windows around boom peaks to examine macro aggregates and firm- and bank-level financial indicators.

### II. Methodology — formal definition and implementation
- Definition:
  - Let lit denote deviation from long-run trend in log real credit per capita for country i at date t; let (σi(l)) denote the standard deviation of this cyclical component.
  - A credit boom is a contiguous period where lit ≥ φ·(σi(l)).
  - Peak date t̂ is where lit − φ·(σi(l)) is maximal within contiguous dates satisfying the boom condition.
  - Start date ts minimizes |lit − φs·(σi(l))| for ts < t̂; end date te minimizes |lit − φe·(σi(l))| for te > t̂.
  - Trend computed with Hodrick-Prescott (HP) filter with smoothing parameter set at 100 (standard for annual data).
- Data construction:
  - Credit measure is the sum of claims on the private sector by deposit money banks (IFS line 22d) plus, when available for the entire sample period for a country, claims by other financial institutions (IFS line 42d).
  - Real credit per capita calculated as the average of two contiguous end-of-year observations of nominal credit per capita deflated by their respective end-of-year CPI.
  - HP smoothing parameter = 100 for detrending; sensitivity with other parameters examined.

### III. Credit boom episodes and main features (empirical counts and stylized facts)
- Counts:
  - Identified 49 credit booms: 27 in industrial countries and 22 in emerging economies.
  - Four emerging economies (India, South Africa, Turkey, and Venezuela) had booms identified in 2006 but were excluded from event analysis because they were not completed.
- Magnitude and relative size:
  - At boom peaks, average expansion in real credit per capita reached almost 30 percent above trend in emerging economies, twice what is observed in industrial countries.
  - In multiples of country-specific standard deviations, booms in industrial and emerging economies are similar (emerging economies exhibit larger cyclical variability).
- Duration and phases:
  - Medians indicate EMs and ICs show booms with similar duration of about 6-7 years and upswings that last longer than downswings.
  - Using means, EMs appear to show longer and more asymmetric booms (driven largely by Sudden Stops cases).
- Geography and synchronization:
  - 40 percent of EM booms observed in East Asia; 32 percent in Latin America.
  - 33 percent of IC booms observed in the G7; 18 percent in the Nordic countries.
  - Booms tend to be synchronized internationally and centered on large events: Bretton Woods collapse (early 1970s), petro dollars prelude to the 1980s debt crisis, ERM and Nordic crises of early 1990s, and recent Sudden Stops.
- Data coverage caveat:
  - Decline in frequency of observed booms in industrial countries may reflect financial deepening away from banks into non-bank financial intermediaries not captured in IFS bank-centered data.

### IV. Differences with Gourinchas, Valdes and Landerretche (GVL) method
- Three principal methodological differences:
  1. Measure of credit: real credit per capita (this paper) v. credit-to-GDP ratio (GVL).
  2. Detrending: standard HP filter over full sample with smoothing parameter 100 (this paper) v. expanding HP trend with smoothing parameter 1000 (GVL).
  3. Thresholds: country-specific thresholds proportional to σi(l) (this paper) v. a boom threshold invariant across countries (GVL).
- Quantitative implications (illustrated with Chile):
  - Different credit measures and smoothing parameters affect boom timing, peak magnitude, and duration.
  - GVL’s higher smoothing parameter (1000) produces a smoother trend, larger and more persistent deviations, and can treat high observed credit as part of trend (lengthening booms).
  - Using country-specific thresholds ensures robustness to credit measure choice; GVL’s invariant threshold may misclassify booms relative to country cyclical variability.

### V. Credit booms and macroeconomic dynamics — event analysis (seven-year windows)
- Variables analyzed (detrended with HP λ=100 unless noted): output (Y), private consumption (C), public consumption (G), investment (I), nontradables output (YN), real exchange rate (RER), current account-output ratio (CAY), capital inflows share of output (KI). RER not per-capita; CAY and KI in current prices and not per-capita.
- Emerging economies (typical patterns, magnitudes and timing):
  - Y, C and G: rise 2 to 4 percentage points above trend in build-up; fall 3 to 4 percent below trend in downswing.
  - Investment I: rises up to 18 percent above trend at t = -1 and drops below trend by a similar amount by t = 2.
  - Nontradables YN: rises to about 6½ percent above trend by t = 0 and drops to almost 3 percent below trend by t = 3.
  - Median RER: appreciates 9 percent above trend at t = -1, drops to about 4 percent below trend when boom unwinds.
  - CAY: declines to a deficit of about 2½ percentage points of GDP in expansion and rises to a surplus of 1½ percentage points of GDP in decline.
  - KI: rises by up to 3½ percentage points of GDP by t = -1 and then drops by 2¼ percentage points of GDP by t = 3.
- Industrial countries:
  - Similar cyclical pattern but smaller amplitudes.
  - Government consumption shows slight decline in expansion and slight rise in contraction (opposite of EMs).
  - Nontradables sector fluctuations much smaller; RER slightly depreciates in expansions rather than appreciations seen in EMs.
- Statistical significance:
  - Cross-section regressions show majority of mean and median estimates for Y, YN, C, and I are statistically significant at the 1 percent level for both ICs and EMs; G, RER and CA/Y often have larger standard errors.
- Comparison with output booms:
  - Conditioning on output booms (instead of credit booms) yields little discernible credit cycle in EMs; in ICs there is a credit cycle but much less pronounced than during credit booms. Some boom-bust features (e.g., housing price declines) are not present around output booms.
- Prices and asset markets:
  - Industrial countries: negligible inflation changes; equity and housing prices rise in build-up and fall in decline.
  - Emerging economies: mean inflation spikes after boom peaks driven by outliers; median inflation rises only 2 percent above trend. Equity and housing price dynamics similar to ICs but larger in magnitude in EMs.
- Regional heterogeneity:
  - Industrial country variation: Nordic countries show larger fluctuations in credit and macro variables than G7; some macro variables peak earlier than credit in Nordics.
  - Emerging economy variation: credit expansions and consumption fluctuations larger in Latin America; current account reversals larger in Asia.

### VI. Frequency analysis — association with crises and preconditions
- Credit booms and crises (Table 5 findings):
  - Emerging economies: about 68 percent of credit booms associated with currency crises; 55 percent with banking crises; 32 percent with Sudden Stops. Most crises preceded or coincided with boom peaks.
  - Industrial countries: credit booms only occasionally associated with banking and currency crises; no association with Sudden Stops. Combined frequency of currency crises before/after/at peak slightly over 50 percent; banking crises 15 percent.
  - Countries experiencing crises display larger macro fluctuations and more abrupt declines than non-crisis countries.
- Preconditions examined (capital inflows, financial reforms, TFP gains):
  - Large capital inflows defined as preceding three-year average of KI ranking in top quartile of respective country group over 1975-2006.
  - Financial reforms measured with Abiad, Detragiache, and Tressel (2007) index; significant reforms when preceding three-year change ranks in top quartile over 1975-2002.
  - High TFP defined when preceding three-year average of TFP growth ranks in top quartile over 1975-2006.
- Frequency results (Table 6 summary):
  - Industrial countries: 40 percent of credit booms followed large TFP gains; 33 percent followed significant financial reforms; capital inflows less important.
  - Emerging economies: over 50 percent of credit booms preceded by large capital inflows; TFP gains and financial reforms play small roles.
  - Interpretation: IC booms more often follow TFP gains and domestic financial reforms; EM booms more often follow capital inflow surges.
- Exchange rate regimes (Table 7 summary):
  - About ¾ of credit booms occur in countries with managed or fixed exchange rate regimes (applies to ICs, EMs, and all countries combined), using Reinhart and Rogoff (2004) classification.
- Early warning probability (starting threshold crossings):
  - Starting thresholds considered: ½ and 1 standard deviation of cyclical component of credit.
  - Once starting threshold of 1 standard deviation is crossed:
    - Probability of a credit boom = 16 percent for emerging economies, 22 percent for industrial countries, and 18 percent for all countries combined.
  - Once starting threshold of ½ standard deviation is crossed:
    - Probability of a credit boom = 9 percent for emerging economies, 14 percent for industrial countries, and 11 percent for all countries combined.
  - Probabilities are higher for industrial than for emerging economies, indicating greater predictive precision in ICs when the starting threshold is crossed.

### VII. Microeconomic indicators and banking sector patterns (summary of findings referenced)
- During credit booms:
  - Firm-level indicators: leverage, firm values, and use of external financing move procyclically—rising in upswing and declining in downswing.
  - Bank-level indicators: credit issuance and asset returns move procyclically.
  - Bank capital adequacy ratios and non-performing loans move countercyclically—capital ratios fall and non-performing loans rise in aftermath of booms.
- Differences across ICs and EMs:
  - Fluctuations in micro and bank indicators are larger, more persistent, and more asymmetric in EMs, particularly in the nontradables sector.
  - Not all credit booms end in crisis, but many recent EM crises were associated with credit booms.
  - Credit booms in EMs are more frequent when preceded by large capital inflows; in ICs booms are more frequent after high TFP or financial reforms and less frequent after large capital inflows.
- Overall macro-micro concordance:
  - Macro expansions, asset price rises, and external imbalances in build-up phases of booms coincide with firm- and bank-level evidence of increased leverage, risk taking, and bank return expansions; post-peak phases show reversals.

### VIII. Key conclusions and surveillance implications
- The thresholds method with country-specific standard deviation-based thresholds and HP detrending with λ = 100 identifies credit booms that are associated with clear macro and micro cyclical patterns and a sizable association with financial crises in emerging economies.
- Credit booms exhibit:
  - Well-defined expansion (build-up) and contraction (downswing) phases with coherent movements in output, expenditures, asset prices, RER, and external balances.
  - Larger, more persistent, and more asymmetric cycles in emerging economies than in industrial countries.
- Surveillance and policy signals:
  - Cross-country patterns indicate that monitored crossings of ½ or 1 standard deviation starting thresholds have measurable, though modest, predictive power (probabilities reported above) for subsequent credit booms and can be used as early-warning indicators.
  - The strong association between managed/fixed exchange rate regimes and booms (about ¾ of booms) suggests exchange rate regime considerations are relevant when assessing boom risks.
  - The differing preconditions for booms in ICs (TFP gains, financial reforms) versus EMs (capital inflows) imply policy responses and preventive measures should be tailored to country context.

*Source: IMF working paper content (excerpts provided in the supplied PDF content)._*

### Section 3 showed that credit booms are accompanied by a synchronized pattern of economic

### _wp08226 - Section 3 showed that credit booms are accompanied by a synchronized pattern of economic

### Data and event-window construction
- Credit boom events identified in Section 2 are used to construct seven-year event windows centered on boom-peak years.
- Firm-level sample:
  - Source: Worldscope and Datastream.
  - Sample period: 1980 to 2005 for the industrial countries and 1991 to 2005 for the emerging markets.
  - Restriction: analysis limited to credit booms dated after 1991 in the emerging markets.
  - Coverage limitations: corporations listed in stock exchanges; fiscal year reporting.
  - Data corrections: multiple listings assigned to country of primary listing; dead stocks removed; outliers removed (negative market capitalizations or observations in excess of two standard deviations from mean).
- Bank-level sample:
  - Source: Bankscope.
  - Coverage period: 1995-2005.
  - Institutions included: commercial banks, saving banks, cooperative banks, mortgage banks, medium- and long-term credit banks, and bank holding companies.
  - Data corrections: central bank, government and multilateral institutions excluded; outliers removed.
  - Industrial-country bank indicators not reported in main analysis because these countries experienced only three credit booms since 1995.
- Aggregation:
  - Country aggregates constructed as medians of firm or bank indicators.
  - Aggregates for emerging and industrial countries generated as medians of the country aggregates.
  - Sectoral breakdown (firm-level): tradables and nontradables, sector aggregates measured using median firm of the sector.
  - Nontradables sector includes: construction, printing and publishing, recreation, retailers, transportation, utilities, and miscellaneous.
  - Tradables sector includes: aerospace, apparel, automotive, beverages, chemical, electrical, electronics, metals, and all others in the database.

### Firm-level indicators (constructed)
- Seven firm-level financial indicators:
  1. Tobin’s Q measures:
     - Q1 = ratio of market to book value of equity.
     - Q2 = ratio of market value plus total debt to book value.
  2. Effective interest rate (ER) = ratio of total debt service to total debt obligations.
  3. Profitability (PR) = return on assets in percent.
  4. Total leverage:
     - LBV = total debt as a percent of book value of assets.
     - LMV = total debt as a percent of market value of assets.
  5. Short-term leverage (SLMV) = short-term debt as a percent of market value of assets.
  6. Working capital leverage (LWK) = current liabilities in percent of sales.
  7. Rajan-Zingales index (RZ) = 1 minus the ratio of cash flow from operations to capital expenditures (cash flow adjusted for changes in inventories, payables, and receivables).

### Firm-level dynamics during credit booms (key findings)
- Leverage:
  - All leverage ratios rise in the build-up phase and collapse in the declining phase.
  - Amplitude of fluctuations larger in emerging economies than in industrial countries.
  - LMV increases particularly large because it includes equity price declines coincident with credit boom peaks.
  - Specific magnitudes:
    - In emerging economies, LBV rises about 20 percentage points from its minimum in the upswing to the peak (compared with 12 percentage points in industrial countries).
    - LWK rises almost 12 percentage points in emerging economies (compared with 2 percentage points in industrial countries).
- Tobin’s Q and profitability:
  - Corporations in both industrial and emerging economies start expansionary phase from high asset valuation and profitability, which decline at the boom peak and remain depressed afterward.
  - Magnitude of fluctuations much larger in emerging economies.
- Effective interest rate (ER):
  - Emerging markets: ER is low in expansionary phase and then jumps about 500 basis points one year after the booms peak.
  - Industrial countries: ER is significantly more stable and drops slightly after credit booms peak.
- Rajan-Zingales (RZ) index:
  - Corporations more dependent on external financing in the build-up phase than in the downswing.
  - Size of correction in the downswing much larger in emerging economies than in industrial countries.
  - Interpretation: consistent with “creditless recoveries” after Sudden Stops where firms adjust to loss of credit by providing internal financing for operational expenses previously financed externally.
- Regional patterns within emerging economies:
  - Asia vs Latin America:
    - Firms in Asia are significantly more leveraged than in Latin America; leverage ratios fluctuate more sharply in Asia.
    - Tobin’s Q in Asia falls sharply in a continuous decline over the seven-year window; in Latin America it shows an ambiguous pattern and a more modest overall decline.
    - Dependence on external financing falls sharply in both regions in the downswing, with a larger correction in Latin America.
  - Sectoral differences (tradables vs nontradables):
    - Leverage cycles (LBV, LMV, LWK) synchronized with credit booms are significantly larger for firms in the nontradables sector and synchronized with sharp boom-bust cycle of the real exchange rate.
    - RZ index: firms in tradables sector slightly more dependent on outside financing in expansionary phase; dependence falls sharply for firms in both sectors, but significantly more for nontradables firms.
    - Consistent with literature that tradables sector has more stable access to credit than nontradables.

### Bank-level indicators (constructed)
- Four bank-level financial indicators:
  1. Profitability: ROA = ratio of net income to average assets (average assets = mean of end-of-period assets in year t and t-1).
  2. Non-performing loans: NPL as measure of asset quality.
  3. Lending activity: LAR = ratio of bank loans to total assets.
  4. Capital adequacy: CAR = bank’s capital as ratio of its risk-weighted assets.
- Basel benchmark:
  - Minimum CAR recommended by the Basel Committee is 8 percent.
  - Supervisors in more vulnerable economies (such as EMs) are encouraged to set higher minimum CARs (often in the 8-12 percent range).

### Bank-level dynamics during credit booms (emerging economies)
- Lending and asset quality:
  - Lending activity high in expanding phase of credit booms.
  - Non-performing loans rise from 2.5 percent a year before the peak to 10 percent two years after the peak.
  - Interpretation: supports the argument that lowering lending standards in booms increases lending to riskier clients, weakening asset quality.
- Profitability:
  - ROA reaches highest level at 1.5 percent a year before the peak, followed by a collapse in the ending phase.
- Capital adequacy:
  - CAR shows v-shaped dynamics: drops sharply late in the expanding phase and then rebounds early in the ending phase, ending at higher levels than when booms start.
  - CAR trough was just below 12 percent, indicating vulnerabilities in banking systems of these economies.

### Macro-micro synchronization and summary statistics
- Count of identified credit booms (1960-2006):
  - 27 credit booms in industrial countries.
  - 22 credit booms in emerging economies.
- General synchronous pattern across booms:
  - Build-up phase associated with expansions in output, rising equity and housing prices, real currency appreciation, and widening external deficits; downswing shows opposite dynamics.
  - Similar cyclical dynamics in firm-level leverage and valuation, and in bank-level asset quality, profitability and lending activity.
  - Credit booms tend to be synchronized internationally and centered on large events (examples cited: 1980s debt crisis, 1992 ERM crisis, Sudden Stops in emerging economies).
- Differences between industrial and emerging economies:
  1. Fluctuations in macroeconomic aggregates and micro-level indicators are larger, more persistent, and asymmetric in emerging economies; pattern particularly strong in nontradables sector.
  2. Not all credit booms end in crisis, but many recent emerging market crises were associated with credit booms.
  3. Credit booms in emerging economies tend to be preceded by large capital inflows and not by domestic financial reforms or TFP gains; credit booms in industrial countries tend to be preceded by high TFP or financial reforms.
  - Additional finding: frequency of credit booms is higher in economies with fixed or managed exchange rates.
- Tables and event-window statistics (selected reported fractions and means):
  - Table 1 (Credit Booms: Duration) reports mean and median durations and fractions spent in upswing/downturn across thresholds (0.00, 0.25, 0.50, 0.75, 1.00) for Emerging Market Economies and Industrial Countries.
  - Table 2 (Coincidence of Credit Booms with Output and Demand Booms) reports frequencies (fractions of credit booms coinciding within seven-year window):
    - Output: Industrial Countries 0.56; Emerging Market Economies 0.32; All 0.45.
    - Non-tradable Output: Industrial Countries 0.33; Emerging Market Economies 0.55; All 0.43.
    - Consumption: Industrial Countries 0.56; Emerging Market Economies 0.36; All 0.47.
    - Investment: Industrial Countries 0.59; Emerging Market Economies 0.55; All 0.57.
    - Government Expenditures: Industrial Countries 0.33; Emerging Market Economies 0.41; All 0.37.
  - Table 3 provides event-window t-coefficients and standard errors for mean values (real credit, output, non-tradable output, consumption, government consumption, investment, REER, current account balance) across t-3 to t+3 for Industrial Countries and Emerging Market Economies (detailed t-coefficients and standard errors reported in table).

### Policy implications and conclusions
- Methodological:
  - The thresholds method proposed provides a tractable framework for measuring and identifying credit booms associated with cyclical fluctuations in macro aggregates and key financial indicators.
  - Credit booms can be identified by the size of credit expansion relative to trend and supplemented with indicators of excessive credit growth.
- Indicators signaling excessive credit growth (to supplement credit measures):
  - Booms in output and expenditures.
  - Excessive real appreciation and/or expansion of the nontradables sector.
  - Large inflows of foreign capital (in EMs).
  - Fast TFP growth or domestic financial reforms (in ICs).
  - Increases in leverage and profitability ratios of corporations.
  - Weakening in the quality of banks’ assets.
- Policy recommendation:
  - Use of corrective policy actions to prevent credit booms is important because the declining phase of credit booms is associated with recessions and a higher incidence of financial crises.
- Research implication:
  - The empirical regularities constitute stylized facts for models of “credit transmission” that should explain (1) the strong association of credit booms with booms in output and expenditures, rising asset prices, widening external deficits and sharp real appreciations, and (2) the close relationship between these macro features and cyclical patterns in firms’ and banks’ financial indicators.

*Italic: Source — _wp08226 - Section 3 showed that credit booms are accompanied by a synchronized pattern of economic (PDF).*

### 2.  Median Values

### _wp08226 - 2.  Median Values

### Statistical significance of event-window coefficients (Table 3)
- Real credit: coefficients (median values and standard errors in brackets) across event window show:
  - t-3: -0.015 (0.015)
  - t-2: -0.015 (0.015)
  - t-1: -0.015 (0.013)
  - t-0: -0.015 (0.013)
  - t+1: -0.015 (0.013)
  - t+2: -0.015 (0.014)
  - t+3: -0.015 (0.015)
  - subsequent coefficients: 0.010 (0.022), 0.048** (0.023), 0.157*** (0.023), 0.260*** (0.023), 0.212*** (0.025), 0.052** (0.024), -0.019 (0.024)
- Output (Y): coefficients and standard errors:
  - t-3: 0.009 (0.006)
  - t-2: 0.017*** (0.006)
  - t-1: 0.031*** (0.006)
  - t-0: 0.028*** (0.005)
  - t+1: 0.016*** (0.006)
  - t+2: -0.010* (0.006)
  - t+3: -0.015*** (0.006)
  - subsequent: 0.004 (0.009), 0.023** (0.009), 0.027*** (0.009), 0.036*** (0.009), -0.011 (0.01), -0.014* (0.008), -0.032*** (0.008)
- Non-tradable Output (YN): coefficients and standard errors:
  - t-3: 0.007 (0.009)
  - t-2: 0.013 (0.008)
  - t-1: 0.023*** (0.008)
  - t-0: 0.029*** (0.008)
  - t+1: 0.011 (0.008)
  - t+2: -0.004 (0.008)
  - t+3: -0.009 (0.008)
  - subsequent: 0.010 (0.009), 0.054*** (0.008), 0.052*** (0.008), 0.066*** (0.008), 0.013 (0.009), -0.016* (0.009), -0.025*** (0.01)
- Consumption (C) — Emerging Market Economies vs Industrial Countries (event-window coefficients):
  - Emerging Market Economies consumption coefficients:
    - t-3: 0.005 (0.006)
    - t-2: 0.011** (0.006)
    - t-1: 0.031*** (0.006)
    - t-0: 0.022*** (0.005)
    - t+1: 0.009* (0.005)
    - t+2: -0.004 (0.006)
    - t+3: -0.017*** (0.005)
    - subsequent: 0.003 (0.009), 0.035*** (0.009), 0.042*** (0.009), 0.041*** (0.009), -0.009 (0.01), -0.023** (0.01), -0.026*** (0.01)
  - Industrial Countries consumption coefficients (t-3 to t+3):
    - 0.003 (0.009), 0.035*** (0.009), 0.042*** (0.009), 0.041*** (0.009), -0.009 (0.01), -0.023** (0.01), -0.026*** (0.01)
- Government consumption (median coefficients and standard errors):
  - t-3: -0.010** (0.004)
  - t-2: -0.012*** (0.004)
  - t-1: 0.003 (0.005)
  - t-0: 0.009** (0.005)
  - t+1: 0.010** (0.005)
  - t+2: 0.005 (0.005)
  - t+3: 0.006 (0.005)
  - industrial-country sequence: 0.015 (0.012), 0.015 (0.012), 0.029** (0.012), 0.031*** (0.012), 0.017 (0.012), 0.009 (0.012), -0.022* (0.012)
- Investment (I) — median coefficients:
  - Emerging Market Economies sequence examples:
    - 0.025* (0.014), 0.045*** (0.014), 0.111*** (0.016), 0.101*** (0.016), 0.027* (0.015), -0.067*** (0.013), -0.076*** (0.014)
  - Industrial Countries sequence examples:
    - 0.035 (0.027), 0.128*** (0.025), 0.173*** (0.027), 0.169*** (0.026), -0.046* (0.028), -0.134*** (0.029), -0.051* (0.029)
- Real Exchange Rate (REER) — median coefficients:
  - t-3: -0.006 (0.012)
  - t-2: -0.015 (0.013)
  - t-1: 0.001 (0.012)
  - t-0: 0.026* (0.014)
  - t+1: 0.014 (0.012)
  - t+2: -0.011 (0.01)
  - t+3: -0.006 (0.01)
  - industrial-country later coefficients: 0.028 (0.028), 0.056** (0.026), 0.067*** (0.025), 0.020 (0.025), 0.022 (0.026), -0.008 (0.026), -0.038 (0.025)
- Current Account Balance (CAY) — median coefficients:
  - t-3: 0.003 (0.003)
  - t-2: -0.001 (0.003)
  - t-1: -0.009** (0.003)
  - t-0: -0.018*** (0.003)
  - t+1: -0.009*** (0.003)
  - t+2: -0.001 (0.003)
  - t+3: 0.003 (0.003)
  - industrial-country later coefficients: -0.007 (0.005), -0.007 (0.004), -0.024*** (0.005), -0.022*** (0.005), 0.015** (0.006), 0.006 (0.005), 0.002 (0.005)

Note: Standard errors are in brackets. The symbols *, **, and *** indicate statistical significance at the 10%, 5%, and 1% level, respectively. The coefficients are obtained by regressing each macroeconomic aggregate on a constant.

### Regional features — Cross-country median of cyclical components (Table 4)
- Panel A. Industrial Countries: G7 vs Nordic Countries (median cyclical components by date t-3 to t+3)
  - Real credit (G7): t-3: -0.76, t-2: 3.89, t-1: 7.76, t-0: 10.82, t+1: 9.09, t+2: 3.26, t+3: -1.26
  - Real credit (Nordic): t-3: -1.05, t-2: 7.68, t-1: 13.54, t-0: 17.77, t+1: 14.67, t+2: 7.77, t+3: 1.29
  - Output (Y) (G7): t-3: 0.88, t-2: 1.84, t-1: 2.78, t-0: 1.87, t+1: -0.33, t+2: -1.98, t+3: -1.35
  - Output (Y) (Nordic): t-3: 2.07, t-2: 3.55, t-1: 3.89, t-0: -0.49, t+1: -1.48, t+2: -2.64, t+3: -2.34
  - Non-tradables Output (YN) (G7): 1.46, 1.70, 1.80, 1.55, -0.05, -1.64, -0.45
  - Non-tradables Output (YN) (Nordic): 3.11, 4.83, 5.78, 3.75, 0.63, 0.79, -2.74
  - Consumption (C) (G7): 1.04, 2.08, 2.77, 1.77, 0.26, -1.06, -1.76
  - Consumption (C) (Nordic): 4.22, 5.22, 4.92, 2.06, -0.81, -2.81, -2.54
  - Investment (I) (G7): 3.62, 7.67, 9.48, 7.08, -2.07, -7.50, -5.14
  - Investment (I) (Nordic): 4.35, 9.20, 19.64, 6.46, -0.58, -7.61, -12.31
  - Real Exchange Rate (RER) (G7): -2.89, -0.63, 2.87, 3.87, 2.48, 3.75, -1.99
  - Real Exchange Rate (RER) (Nordic): -0.77, -2.46, 0.18, 3.60, 1.44, 0.77, -0.18
  - Current Account-GDP ratio (CAY) (G7): -2.89? (source layout ambiguous for this line) — also shows G7 sequence: 0.09, -0.31, -0.88, -1.09, 0.05, 0.05, 0.06
  - Current Account-GDP ratio (CAY) (Nordic): 0.08, -2.10, -2.71, -1.42, -0.78, -0.31, 0.82
- Panel B. Emerging Economies: Latin America (LA) vs Asia (median cyclical components by date t-3 to t+3)
  - Real credit (LA): 13.22, 29.52, 43.14, 48.14, 39.64, 12.61, -6.30
  - Real credit (Asia): 1.09, 4.90, 12.22, 24.68, 21.10, 1.34, -3.97
  - Output (Y) (LA): 1.10, 5.50, 4.64, 5.73, 1.06, -2.32, -5.19
  - Output (Y) (Asia): 3.78, 3.38, 4.41, 3.55, -3.89, -1.48, -3.19
  - Non-tradables Output (YN) (LA): 0.98, 6.35, 6.02, 6.90, 1.81, 0.28, -2.90
  - Non-tradables Output (YN) (Asia): 4.15, 5.39, 6.77, 6.51, -0.36, -3.08, -2.00
  - Consumption (C) (LA): 0.88, 5.50, 7.21, 8.78, 2.10, -4.65, -5.41
  - Consumption (C) (Asia): 2.99, 3.93, 6.18, 1.12, -5.23, -2.22, -1.56
  - Investment (I) (LA): 8.95, 23.02, 17.84, 20.62, 3.26, -17.89, -22.12
  - Investment (I) (Asia): 12.08, 14.29, 20.70, 17.00, -11.58, -11.13, -4.76
  - Real Exchange Rate (RER) (LA): 10.40, 16.50, 13.71, 16.85, 12.83, -2.36, -3.74
  - Real Exchange Rate (RER) (Asia): 4.22, 7.05, 10.97, 8.96, -6.71, -1.30, -6.88
  - Current Account-GDP ratio (CAY) (LA): -0.84, -1.68, -2.41, -2.96, -0.86, 1.46, 0.24
  - Current Account-GDP ratio (CAY) (Asia): -2.50, -3.24, -3.22, -2.37, 3.40, 0.40, 0.66

### Credit booms and crises — Frequencies (Table 5)
- All countries (frequency in seven-year window around boom):
  - Banking crises: Before 0.15, Peak 0.15, After 0.04, Total 0.33
  - Currency crises: Before 0.23, Peak 0.15, After 0.23, Total 0.60
  - Sudden Stops: Before 0.06, Peak 0.06, After 0.04, Total 0.17
- Industrial Countries:
  - Banking crises: Before 0.08, Peak 0.08, After 0.00, Total 0.15
  - Currency crises: Before 0.15, Peak 0.08, After 0.31, Total 0.54
  - Sudden Stops: Before 0.00, Peak 0.00, After 0.04, Total 0.04
- Emerging Market Economies:
  - Banking crises: Before 0.23, Peak 0.23, After 0.09, Total 0.55
  - Currency crises: Before 0.32, Peak 0.23, After 0.14, Total 0.68
  - Sudden Stops: Before 0.14, Peak 0.14, After 0.05, Total 0.32

Notes:
- 1/ Coincidence of credit booms and financial crises in the seven-year window around the boom.
- 2/ Banking crises as defined by Demirguic-Kunt and Detragiache (2006). See Appendix 2 for details.
- 3/ Currency crises as defined by Eichengreen and Bordo (2002). See Appendix 2 for details.
- 4/ Sudden Stops as defined by Calvo, Izquierdo and Mejia (2004). See Appendix 2 for details.

### Potential triggering factors of credit booms — Frequency distribution (Table 6)
- Frequency (Industrial Countries / Emerging Market Economies / All):
  - Large Capital Inflows (A): 0.27 / 0.50 / 0.52
  - Significant Productivity Gains (B): 0.40 / 0.17 / 0.33
  - Large Financial Sector Changes (C): 0.33 / 0.22 / 0.27
  - Other: 0.27 / 0.39 / 0.30
- Memo items (joint occurrences):
  - (A) & (B): 0.07 / 0.00 / 0.09
  - (A) & (C): 0.00 / 0.06 / 0.12
  - (B) & (C): 0.07 / 0.11 / 0.09
  - (A) & (B) & (C): 0.07 / 0.06 / 0.06

Notes:
- 1/ Because of data availability we have used the 1975-2005 period only.
- 2/ The three-year average of capital inflow before the peak of the boom ranks in the top quartile of their corresponding country group.
- 3/ The three-year average of the annual growth rate of TFP before the peak of the boom ranks in the top quartile of their corresponding country group.
- 4/ The three-year change before the peak of the boom in the financial reform index ranks in the top quartile of their corresponding country group.

### Credit booms and exchange rate regimes (Table 7)
- Frequency distribution by regime (Industrial Countries / Emerging Market Economies / All):
  - Fixed and managed1/: 77.78 / 73.91 / 76.00
  - Dirty Floating2/: 7.41 / 17.39 / 12.00
  - Floating3/: 7.41 / 0.00 / 4.00
  - Mixed: 7.41 / 8.70 / 8.00

Notes:
- 1/ Fixed and managed includes the following regimes from the Reinhart-Rogoff (2004) classification: no separate legal tender, pre-announced peg or currency board arrangement, pre-announced horizontal band that is narrower than or equal to +/- 2%, de facto peg, pre-announced crawling peg, pre-announced crawling band that is narrower than or equal to +/-2%, de facto crawling peg, and de facto crawling band that is narrower than or equal to +/-2%.
- 2/ Dirty floating includes the following regimes from the Reinhart-Rogoff (2004) classification: pre-announced crawling band wider than or equal to +/-2%, de facto crawling band narrower than or equal to +/-5%, moving band that is narrower than or equal to +/- 2%, and managed floating.
- 3/ Freely floating regimes from the Reinhart-Rogoff (2004) classification.

### Sample of countries (Appendix I)
- Industrial countries included (with peak years where identified): Australia (AUS, 1989), Austria (AUT, 1972 and 1980), Belgium (BEL, 1980 and 1990), Canada (CAN, 1967), Denmark (DNK, 1987 and 1991), Finland (FIN, 1991), France (FRA, 1991), Germany (DEU, 1973), Greece (GRC, 1972), Ireland (IRL, 1979 and 2000), Italy (ITA, 1973 and 1992), Japan (JPN, 1973), Netherlands (NLD, 1979), New Zealand (NZL, 1974), Norway (NOR, 1988), Portugal (PRT, 1973 and 2001), Spain (ESP), Sweden (SWE, 1990), Switzerland (CHE, 1990), United Kingdom (GBR, 1974 and 1990), and United States (USA, 1999).
- Emerging Market Economies included (with peak years where identified): Algeria (DZA), Argentina (ARG, 1982), Brazil (BRA), Chile (CHL, 1981), Colombia (COL, 1997), Costa Rica (CRI, 1979), Côte d'Ivoire (CIV, 1978), Ecuador (ECU), Egypt (EGY,1982), Hong Kong (HKG, 1998), India (IND,*), Indonesia (IDN, 1997), Israel (ISR, 1979), Jordan (JOR, 1970), Korea (KOR, 1998), Malaysia (MYS, 1997), Mexico (MEX, 1994), Morocco (MAR, 1977), Nigeria (NGA, 1982), Pakistan (PAK, 1986), Peru (PER), Philippines (PHL, 1983 and 1997), Singapore (SGP, 1984), South Africa (ZAF, *), Thailand (THA, 1997), Turkey (TUR, 1997 and *), Uruguay (URY, 1981), and Venezuela, Rep. Bol. (VEN, *).
- (*) Ongoing credit booms.

### Data definitions and sources (Appendix II)
- A. Macroeconomic and financial data (variable definitions and data sources):
  - Credit to the non-financial private sector: Sum of claims on the private sector by deposit money banks (IFS line 22d) plus, whenever available for the entire sample period by other financial institutions (IFS line 42d). Data sources: IFS, OECD Analytic Database, Datastream, Haever Analytics.
  - Consumer price index: Consumer price index (both average and end-of-period). Source: IFS.
  - Nominal GDP: GDP in current prices, local currency. Source: WDI.
  - Population: Population. Source: WDI.
  - Real GDP: Real GDP per-capita, in international prices. Source: PWT 6.2.
  - Private consumption: Real private consumption per-capita, in international prices. Source: PWT 6.2.
  - Government consumption: Real government consumption per-capita, in international prices. Source: PWT 6.2.
  - Investment: Real investment per-capita, in international prices. Source: PWT 6.2.
  - Non-tradable GDP: Sum of the value added in services plus the value added in industry minus manufacture. Source: WDI.
  - Current account balance: Current account balance as percent of GDP. Source: WDI.
  - Real exchange rate: Real effective exchange rate, index. Source: INS (IMF).
  - Capital inflows: Capital inflows (proxied as the flow of total external liabilities) as percent of GDP. Source: IFS.
  - Real stock prices: Equity price indices deflated using consumer price indeces. Authors' calculation with data from IFS.
  - Real house prices: House price indices deflated using consumer price indeces. Authors' calculation with data from several country sources, Haever Analytics and Bloomberg.
  - Total factor productivity: Total factor productivity calculated using the PWT6.2 dataset. Source cited: Kose, Prasad, and Terrones (2008).
- B. Microeconomic and financial data (definitions and sources):
  - Tobin's Q: Measured by (1) ratio of market value of a firm's equity to the current replacement cost of assets; (2) ratio of market value of a firm's equity and debt to the current replacement cost of assets. Source: Authors' calculations with data from Worldscope.
  - Leverage: Measured by several ratios (total debt to book value of equity; total debt to market value of equity; short term debt to market value of equity; current liabilities to sales). Source: Authors' calculations with data from Worldscope.
  - Effective interest rate, Return on assets, Rajan-Zingales index: Authors' calculations with data from Worldscope.
  - Non-performing loans to gross loans ratio, Loan to asset ratio, Return on average assets, Capital adequacy ratio: Authors' calculations with data from Bankscope.
- C. Crises definitions:
  - Banking Crises: A situation in which at least one of the following conditions holds: (1) the ratio of non-performing assets to total assets of the banking sector exceeds 10 percent; (2) the cost of banking system bailouts exceeds 2 percent of GDP; (3) there is a large scale bank nationalization as result of banking sector problems; and (4) there are bank runs or new important depositor protection measures. Source: Demirguic-Kunt and Detragiache (2006).
  - Currency Crises: A situation in which a country experiences a forced change in parity, abandons a currency peg or receives a bailout from an international organization, and at the same time an index of exchange rate market pressure (a weighted average of the depreciation rate, change in short-term interest rate, and percentage change in reserves) rises. Source: Eichengreen and Bordo (2002).

*Source: _wp08226 - 2.  Median Values (PDF).*

### 1.5 standard deviations above its mean.

### _wp08226 - 1.5 standard deviations above its mean.

### Definitions and event identification
- Sudden Stops: a situation in which a country experiences a year-on-year fall in capital flows that exceeds 2 standard deviations relative to the mean.
- Credit boom (GVL definition): contiguous dates where the deviation from the expanding trend in the credit-GDP ratio satisfies the relative deviation threshold φ: (EHP_it/LY_it) - (EHP_it/LY_it) ≥ φ (GVL notation as given in source).
- Peak of a credit boom: the date within the identified contiguous set with the largest deviation from the expanding HP trend.
- Starting (ending) date of a boom: the date earlier (later) than the peak at which the credit-GDP ratio is higher (lower) than a “limit threshold.”

### Financial reform index
- The index captures changes in seven financial policy dimensions:
  1. credit controls and reserve requirements;
  2. Interest rate controls;
  3. Entry barriers;
  4. State ownership in the banking sector;
  5. Capital account restrictions;
  6. Prudential regulations and supervision of the banking sector;
  7. Securities market policy.
- Construction: the index is the sum of these seven dimensions (each of which can take values between 0 and 3).
- Range: takes values between 0 (the lowest) and 21 (the highest).

### Gourinchas-Valdes-Landerretche (GVL) thresholds method — credit measure and trend
- Credit measure: ratio of nominal credit to the private sector to nominal GDP (LY_it/LY_it notation in source).
- Trend specification: an “expanding trend” using the HP filter.
  - Given dates t = 1,...,T, EHP_it/LY denotes the expanding HP trend of the credit-GDP ratio for country i at date t.
  - The expanding trend is the trend component of the HP filter applied to sequences that expand each year by including data from an initial window of length k.
- GVL implementation details as used in the source:
  - sample for 1960-1996;
  - set k = 5, so the expanding trend starts in 1965;
  - EHP_i,1965/LY is the 1965 value of the HP trend computed using data from 1960 to 1965;
  - HP smoothing parameter set to 1000.
- Interpretation and critique noted in the source:
  - The expanding trend is justified by GVL as reflecting information available to policymakers at a given time.
  - The source comments that the expanding trend yields a trend process close to a smoothed, one-period lagged transformation of the original credit series, which may lead policymakers to misinterpret recent credit observations as part of a trend rather than a boom.
  - Credit data revisions complicate the claim that the expanding trend represents information available to policymakers.

### Thresholds and parameter values used in GVL event analysis
- Relative deviation threshold φ = 19.5 percent.
- Limit threshold = 5 percent.
- Note: GVL also used an “absolute deviation threshold” (measuring credit booms relative to size of the economy) but their macro event analysis is based on the relative threshold.

### Additional methodological notes
- Since credit is a stock variable, GVL proxy Y_it as the geometric average of nominal GDP in t and t+1.
- The source notes that the limit threshold does not correct for the discrete nature of the data; the authors of the source defined starting and ending thresholds using an absolute-value metric to address discreteness.

*Italicized source attribution provided by the pipeline.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp08226.pdf_