## _wp14155

## Source details

**Canonical URL:** [_wp14155](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp14155.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp14155.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp14155.pdf.json)

---

### Introduction and motivation
- Macro-prudential policies aim to contain (the buildup of) systemic risks and achieve greater financial stability, complementing micro-prudential regulation and macroeconomic management tools (monetary and fiscal policies).
- Emerging markets have had much greater experiences with macro-prudential policies, driven by exposures to volatile international capital flows, commodity price shocks, and other risks.
- Research questions:
  - What macro-prudential policies are available in principle and what policies have countries actually used?
  - What is the evidence to date on the effectiveness of these different policies?
  - What are the specific experiences with policies in terms of reducing banking systems’ vulnerability?
- Novelty: extends prior aggregate-level studies by investigating how policies affect behavior at the microeconomic level—specifically, the buildup of vulnerability in individual banks’ balance sheets.

### Data and empirical scope
- Sample and coverage:
  - some 18,000 observations on approximately 2,820 banks over 2000-2010.
  - approximately 1,650 banks in 23 advanced countries.
  - approximately 1,170 banks in 25 emerging markets.
  - 48 countries in total; 35 countries implemented at least one instrument at least once during 2000-2010; 13 countries never used any instrument in this period.
- Bank-level dataset construction:
  - Bank balance sheet data from Bankscope (annual, in US dollars, unconsolidated).
  - Top 100 banks per country where possible.
  - Winsorization: discard observations above/below the five percent level in both tails.
  - Final bank-level coverage: some 18,000 observations on 2,820 banks in 48 countries over 2000-2010.
- Primary dependent variable: change in banks’ total assets (asset growth).
- Additional metrics reported:
  - Asset growth after winsorization: mean 13.32 percent per year; range -37.55 to 77.05.
  - Leverage ratio — Obs. 2,122; Mean = 13.55; Std. Dev. = 10.55; Min = 0.84; Max = 39.70.
  - Loan-to-deposit ratio — Obs. 1,816; Mean = 1.47; Std. Dev. = 1.56; Min = 0.15; Max = 6.36.
  - Bank assets (USD $M) — Obs. 2,123; Mean = 458030.42; Std. Dev. = 1151717.10; Min = 0.01; Max = 3914824.
  - Country-level real GDP growth (%) — Obs. 2,821; Mean = 3.08; Std. Dev. = 3.10; Min = -3.31; Max = 9.52.
  - Interest rate change — Obs. 2,277; Mean = 5.34; Std. Dev. = 17.42; Min = -25.80; Max = 43.51.
- Econometric approach:
  - Generalized Method of Moments (GMM) panel regressions with lagged dependent variables (up to lag 4) and instruments to address endogeneity.
  - Policy variables entered as dummies and interactions with lagged ∆Y to capture intensity effects.
  - Phase analysis: upswing vs downswing determined by country real credit increase/decrease in the year; regressions include interaction of policy dummies with downswing dummy.

### Macro-prudential policy taxonomy used in analysis
- Groupings and instruments:
  - Borrower-targeted measures:
    - caps on debt-to-income (DTI) ratios
    - caps on loan-to-value (LTV) ratios
  - Banks’ assets or liabilities measures:
    - limits on credit growth (CG)
    - foreign currency credit growth (FC)
    - reserve requirements (RR)
  - Counter-cyclical buffers:
    - counter-cyclical capital (CTC)
    - dynamic provisioning (DP)
    - profits distribution restrictions (PRD)
  - Miscellaneous:
    - category labelled Other (institutional measures, accounting, compensation, taxation/levies)
- Conceptual mapping:
  - Category (a) borrower restrictions mainly affects demand for financing.
  - Categories (b)–(d) mainly affect supply (banks’ balance sheets, buffers, institutional measures).

### Actual use statistics (preserve exact sample counts and frequencies)
- Instrument coverage (total countries that used at least once during 2000-2010):
  - LTV: 24 countries; Frequency of Use: 44%; Emerging Markets: 15; Advanced Countries: 9; Frequency of EMs-year: 35%; Frequency of ACs-year: 74%.
  - DTI: 7 countries; Frequency of Use: 9%; Emerging Markets: 5; Advanced Countries: 2; Frequency of EMs-year: 8%; Frequency of ACs-year: 11%.
  - CG: 6 countries; Frequency of Use: 8%; Emerging Markets: 5; Advanced Countries: 1; Frequency of EMs-year: 10%; Frequency of ACs-year: 1%.
  - FC: 8 countries; Frequency of Use: 8%; Emerging Markets: 7; Advanced Countries: 1; Frequency of EMs-year: 10%; Frequency of ACs-year: 3%.
  - RR: 5 countries; Frequency of Use: 5%; Emerging Markets: 5; Advanced Countries: 0; Frequency of EMs-year: 7%; Frequency of ACs-year: 0%.
  - DP: 9 countries; Frequency of Use: 9%; Emerging Markets: 8; Advanced Countries: 1; Frequency of EMs-year: 9%; Frequency of ACs-year: 11%.
  - CTC: 2 countries; Frequency of Use: 1%; Emerging Markets: 2; Advanced Countries: 0; Frequency of EMs-year: 0%; Frequency of ACs-year: 0%.
  - PRD: 6 countries; Frequency of Use: 6%; Emerging Markets: 6; Advanced Countries: 0; Frequency of EMs-year: 4%; Frequency of ACs-year: 0%.
  - Other MaPP: 13 countries; Frequency of Use: 12%; Emerging Markets: 12; Advanced Countries: 1; Frequency of EMs-year: 15%; Frequency of ACs-year: 1%.
- Usage intensity (country-year weighting):
  - LTV: about 44% of policy country-year observations.
  - DTI, CG, FC, DP: about 8% each.
  - RR: 5%; PRD: 3%; CTC: 1%.
- Cross-country patterns:
  - Emerging markets used macro-prudential policies more than advanced countries; macro-prudential policies were four times more likely to be used by emerging markets than advanced countries in the period right before the crisis; this ratio declined to 3.3 after the crisis.
  - Closed capital account countries (Chinn-Ito 2005 below global median) used macro-prudential policies somewhat more than open ones.
  - Advanced-country usage is concentrated in LTVs (74% of advanced countries’ usage by country-year observations).

### Main empirical findings on effectiveness — summary
- Overall effectiveness:
  - Borrower-targeted measures are effective in (indirectly) reducing the buildup of banking system vulnerabilities.
  - Measures aimed at banks’ assets and liabilities are very effective.
  - Counter-cyclical buffers as a group show less promise.
  - The category Other is very effective.
- Phase-specific results:
  - During upswings:
    - All policy sets except the buffer-based category directly help reduce asset growth.
    - Policies aimed at banks’ assets and liabilities and Other measures help the most.
    - Borrower-based measures help reduce asset growth to a lesser degree.
  - During downswings:
    - Borrower-based measures stop declines in bank asset growth in a statistically significant way.
    - Measures aimed at banks’ assets and liabilities and Other have positive impact in contractionary periods, but significance levels are only 25% and 16% respectively in some regressions.
    - Buffer-building measures are not productive in downswings (or upswings).
- Interpretation and mechanisms:
  - Asset and liability-based measures operate directly on banks’ balance sheets, explaining their strong effect on asset growth.
  - Demand-oriented measures (largely aimed at real-estate markets) are effective in addressing banking vulnerabilities; real estate cycles often trigger systemic risk concerns and are closely monitored, potentially reducing circumvention.
  - Some macro-prudential policies can work perversely during downturns if not relaxed sufficiently and in a timely manner—potentially exacerbating downturns.

### Regression results — selected exact coefficient highlights (Table 5 and related tables)
- Regression sample sizes:
  - Base regressions: about 2,630 banks and 13,804 observations in column (1).
  - Subsequent regressions: about 1,667 banks and 8,527 observations in columns (2)-(6).
- Key control coefficients (selected, exact values from Table 5):
  - Lagged Asset Growth:
    - Column (1): -0.346*** [0.051]
    - Column (2): -0.322*** [0.106]
    - Column (3): -0.468** [0.221]
  - Lagged Real GDP Growth (%):
    - Column (1): 4.106*** [1.033]
    - Column (2): 7.705*** [2.829]
    - Column (3): 7.319 [5.022]
  - Lagged Interest Rate Change:
    - Column (1): 0.207** [0.090]
    - Column (2): 0.186 [0.123]
    - Column (3): 0.019 [0.229]
  - Lagged Leverage Ratio:
    - Column (1): -0.022* [0.012]
    - Column (2): -0.029** [0.015]
    - Column (3): -0.024 [0.020]
- MaPP group and individual policy coefficients (selected exact values):
  - Subgroup2 (financial institutions’ asset/liability measures) — Column (1): -0.656*** [0.183]; Column (2): -0.631*** [0.191]; Column (3): -0.994*** [0.344].
  - LTV:
    - Column (3): -0.852*** [0.225]
  - DTI:
    - Column (3): 1.297 [1.126]
    - Column (5): 6.009* [3.128]
  - CG:
    - Column (3): -0.704** [0.346]
    - Column (5): -2.656* [1.532]
  - FC:
    - Column (3): -0.392 [0.244]
    - Column (5): -2.335*** [0.857]
  - CTC:
    - Column (3): -0.406*** [0.099]
    - Column (5): 0.585 [0.521]
  - Other:
    - Column (3): -0.673*** [0.105]
    - Column (5): -1.443** [0.592]
- Phase-specific results (Table 6, selected exact entries):
  - Downswing dummy: -0.230*** [,0.059] in one specification; -0.141*** [,0.054]; -0.203*** [,0.049]; -0.122** [,0.049] across four columns.
  - Subgroup1 (borrower measures) × downswing: 0.237* [,0.121] (positive interaction indicating borrower-based policies mitigate declines in downswings in that specification).
  - Subgroup2 × downswing: 0.103 [,0.089] (not significant).
  - Subgroup4 × downswing: 0.188 [,0.134] (not significant).
- Emerging vs advanced interactions (Table 7, selected exact entries):
  - Subgroup1 (borrower-based) main effect: -1.274*** [,0.379].
  - Sub1 X Emerging: 1.286** [,0.501] (offsetting effect in emerging markets).
  - Subgroup4 (Other) main effect: -1.008*** [,0.337].
  - Sub4 X Emerging: 0.380 [,0.355] (not statistically significant in that column).
  - Lagged Real GDP Growth in one specification: 15.989*** [,5.156].

### Policy implications and recommendations (based on empirical findings)
- Effective elements of the toolkit:
  - Borrower-based measures (LTV, DTI) for dampening excessive lending growth and leverage in booms.
  - Balance-sheet oriented measures (limits on credit growth (CG), foreign currency lending (FC), reserve requirements (RR)) to constrain asset growth.
  - Other institutional/information/taxation measures can be effective and are frequently used in emerging markets.
- Design and calibration:
  - Policies need to be properly chosen and carefully calibrated to country and financial system characteristics (financial structure, state-owned bank presence, exchange rate regime, degree of financial openness).
  - Advanced economies may derive more benefit from borrower-based tools (reflecting real estate driven cycles).
  - Emerging markets may benefit from packages of policies and coordination with capital flow management tools.
- Timing and adjustment:
  - Policies must be adjusted quickly as circumstances change, including timely relaxation in downturns to avoid exacerbating contractions.
  - Some balance-sheet-based measures may act perversely during downswings if not adjusted in a timely manner.
- Potential costs and risks:
  - These policies imply costs as they affect resource allocations and may limit financial sector development.
  - Poorly designed or wrongly implemented tools can be circumvented, imply further distortions, or work perversely.
  - Risk migration to less-regulated sectors (shadow banking) can offset bank-sector gains; borrower-targeted measures are less easily avoided than bank-targeted ones.
- Institutional and data needs:
  - Stress tests can complement macro-prudential tools by being forward-looking and more granular.
  - Improved data, supervisory capacity, and political/institutional frameworks are important for effective policy implementation.

### Limitations and suggestions for further research
- Econometric and identification caveats:
  - Residual selection, endogeneity, and omitted variables may remain; suggested methods include propensity scoring, matching banks across regimes, and studying subsidiaries of the same foreign bank operating under different regimes.
- Circumvention and system-wide effects:
  - Need to study risk migration to shadow banking and overall systemic risk measures (aggregate credit, asset prices, systemic risk rankings such as Marginal Expected Shortfall or CoVaR).
- Additional empirical extensions:
  - Use of more aggregate or market-based measures (credit spreads, systemic risk rankings) to capture comprehensive effects and circumvention.
  - Longer time-series and more detailed intra-year policy change data (data constraints limited intra-year analysis here).
- Broader assessment:
  - Further study needed on macroeconomic costs/benefits of macro-prudential policies, including impacts on resource allocation, growth, and financial development.

*Source: IMF working paper section — "The Macro‑Prudential Toolkit" and empirical analysis (sample 2000-2010; bank-level GMM regressions using Bankscope; summary of Tables 2–7 and accompanying discussion).*

### 1.  The Macro-Prudential Toolkit .......................................................................................

### 1. The Macro-Prudential Toolkit

### Introduction and motivation
- Recent events have highlighted the high costs of financial crises and the potential economic costs arising from the way financial systems operate.
- Macro-prudential policies aim to contain (the buildup of) systemic risks and achieve greater financial stability, complementing micro-prudential regulation and macroeconomic management tools (monetary and fiscal policies).
- Emerging markets have had much greater experiences with macro-prudential policies, due in part to more pronounced business and financial cycles driven by exposures to volatile international capital flows, commodity price shocks, and other risks.
- Cross-country analysis can help ensure macro-prudential policies are properly designed and calibrated to country characteristics and circumstances.

### Research questions and approach
- The paper asks:
  - What macro-prudential policies are available in principle and what policies have countries actually used?
  - What is the evidence to date on the effectiveness of these different policies?
  - What are the specific experiences with policies in terms of reducing banking systems’ vulnerability?
- Novelty: extends prior aggregate-level studies by investigating how policies affect behavior at the microeconomic level—specifically, the buildup of vulnerability in individual banks’ balance sheets.

### Data and empirical scope
- Sample: some 18,000 observations on approximately 2,820 banks over the period 2000-2010.
- Sample composition:
  - approximately 1,650 banks in 23 advanced countries
  - approximately 1,170 banks in 25 emerging markets
- Advantages of bank-level analysis: greater control for characteristics driving balance sheet behavior and reduced concern for endogeneity (macro-prudential policies are adopted in response to aggregate bank behavior and less likely to individual bank behavior alone).
- Distinctions explored:
  - country circumstances (advanced vs. emerging markets; relatively open and closed capital account economies)
  - phases of the financial cycle (upswing vs. downswing in credit to the private sector)

### Macro-prudential policy groupings (as used in analysis)
- Borrower-targeted measures:
  - caps on debt-to-income (DTI) ratios
  - caps on loan-to-value (LTV) ratios
- Banks’ assets or liabilities measures:
  - limits on credit growth (CG)
  - foreign currency credit growth (FC)
  - reserve requirements (RR)
- Counter-cyclical buffers:
  - counter-cyclical capital (CTC)
  - dynamic provisioning (DP)
  - profits distribution restrictions (PRD)
- Miscellaneous:
  - category labelled Other (some overlap with the first three groups)

### Main findings on effectiveness
- Overall:
  - Policies aimed at borrowers are effective in (indirectly) reducing the buildup of banking system vulnerabilities.
  - Measures aimed at banks’ assets and liabilities are very effective.
  - Counter-cyclical buffers as a group show less promise.
  - The category Other is very effective.
- By phase of the credit cycle:
  - During upswings:
    - All policy sets except the buffer-based category directly help reduce asset growth.
    - Policies aimed at banks’ assets and liabilities and Other measures help the most.
    - Borrower-based measures help reduce asset growth to a lesser degree.
  - During downswings (contractionary periods):
    - Borrower-based measures stop declines in bank asset growth in a statistically significant way.
    - Measures aimed at banks’ assets and liabilities as a group and the category Other have positive impact in contractionary periods, but their significance levels are only 25% and 16% respectively.
    - Measures aimed at building banks’ buffers are not productive in downswings (or upswings).
- Interpretation and mechanisms:
  - Asset and liability-based measures operate directly on banks’ balance sheets, explaining their strong effect on asset growth.
  - Demand-oriented measures (largely aimed at real-estate markets) are effective in addressing banking vulnerabilities; this is relevant because real estate cycles often trigger systemic risk concerns and are generally closely monitored and regulated, potentially reducing circumvention.
  - Some macro-prudential policies aimed at mitigating the buildup of financial vulnerabilities can work perversely during downturns if not relaxed sufficiently and in a timely manner—potentially exacerbating downturns.

### Policy implications and recommendations
- Macro-prudential policies can be important elements in mitigating systemic risk, especially for countries exposed to international shocks.
- Costs and risks:
  - These policies imply costs as they affect resource allocations and may limit (efficient) financial sector development.
  - Poorly designed or wrongly implemented tools can be circumvented, imply further distortions, or even work perversely.
- Key lessons:
  - Policies need to be properly chosen and carefully calibrated to country and financial system characteristics.
  - Policies must be adjusted quickly as circumstances change, including timely relaxation in downturns to avoid exacerbating contractions.

*Source: 1. The Macro-Prudential Toolkit (excerpt).*

### Section 3 presents the data used and the results of the empirical analysis on the effectiveness of

### _wp14155 - Section 3 presents the data used and the results of the empirical analysis on the effectiveness of

### Toolkit and conceptual framework
- Macro-prudential policies are classified in a 3 (goals) × 4 (targets/methods) matrix:
  - Rows (goals): dampen expansionary cyclical phases; reinforce resilience in adverse phases; address interconnectedness/internalize spillovers.
  - Columns (targets/methods): a) quantitative restrictions on borrowers/instruments/activities; b) quantitative restrictions on financial institutions’ balance sheets; c) capital and provisioning requirements; d) other (institutional measures, accounting, compensation, taxation/levies).
- Many micro-prudential tools can be repurposed for macro-prudential objectives by making them time-, institution-, or state-varying.
- Category a) (borrower restrictions) mainly affects demand for financing; categories b)–d) mainly affect supply.
- Typical mapping of instruments:
  - borrower-based: caps on LTV and DTI;
  - asset/liability-based: limits on credit growth (CG), foreign currency lending (FC), reserve requirements (RR);
  - buffer-based: dynamic provisioning (DP), counter-cyclical capital (CTC), limits on profit redistribution (PRD);
  - other: residual institutional/taxation/measurement measures.

### Actual use of macro-prudential policies (sample statistics)
- Data source and coding:
  - Policy usage coded as dummies (1 if used in country-year, 0 otherwise).
  - Nine instruments covered: LTV, DTI, CG, FC, RR, DP, CTC, PRD, Other.
- Sample coverage and key counts:
  - Sample period: 2000-2010.
  - 48 countries (both advanced and emerging markets).
  - 35 countries implemented at least one instrument at least once during 2000-2010; 13 countries never used any instrument in this period.
  - LTV used by 24 countries at least once.
  - DP used by 9 countries; FC by 8; DTI by 7; CG and PRD by 6 each; RR by 5; CTC by 2.
- Usage intensity (weighting by country-year where a policy was used):
  - LTV: about 44% of policy country-year observations.
  - DTI, CG, FC, DP: about 8% each.
  - RR: 5%; PRD: 3%; CTC: 1%.
- Cross-country patterns:
  - Emerging markets used macro-prudential policies more than advanced countries; macro-prudential policies were four times more likely to be used by emerging markets than advanced countries in the period right before the crisis; this ratio declined to 3.3 after the crisis.
  - Closed capital account countries (Chinn-Ito 2005 below global median) used macro-prudential policies somewhat more than open ones.
  - Advanced-country usage is concentrated in LTVs (74% of advanced countries’ usage by country-year observations), with limited use of other tools.

### Literature summary on effectiveness
- Cross-country and micro studies generally find evidence that some policies can mitigate procyclicality of credit and leverage:
  - Lim et al. (2011): LTV, DTI, CG caps, RR, DP mitigate procyclicality of credit.
  - IMF (2012b): capital requirements and RRs affect credit growth; LTV limits and capital requirements affect house price appreciation; RRs reduce portfolio inflows in emerging markets with floating exchange rates.
  - Crowe et al. (2011): maximum LTVs linked to real estate cycle appear effective at curbing booms.
  - Dell’Ariccia et al. (2012): macro-prudential policies reduce incidence of credit booms and probability that booms end badly.
  - Case studies (Jiménez et al., Igan and Kang, Aiyar et al.) show dynamic provisioning, LTV/DTI limits, and bank-specific capital requirements can dampen cycles or bank lending growth.
- Caveats noted in the literature: scope for circumvention, heterogeneous impacts by country structure, and mixed evidence on macroeconomic output effects.

### Empirical analysis: data and methods
- Bank-level dataset construction:
  - Bank balance sheet data from Bankscope (annual, in US dollars, unconsolidated).
  - Sample: top 100 banks per country where possible.
  - Winsorization: discard observations above/below the five percent level in both tails.
  - Final bank-level coverage: some 18,000 observations on 2,820 banks in 48 countries over 2000-2010.
  - Advanced vs emerging split: 1,609 banks in 23 advanced countries; 1,212 banks in 25 emerging markets.
- Dependent variables and robustness checks:
  - Primary dependent variable: change in banks’ total assets (asset growth).
  - Also analyzed: leverage (total assets to total equity) and noncore-to-core liabilities (no strong results for these).
  - Asset growth distribution after winsorization: mean 13% per year; range -38% to 77%.
  - Bank size range in sample: $10 million to $3.9 trillion.
  - Average leverage: 14 (some banks as low as debt-to-equity ratio 1; highest 40).
  - Average loan-to-deposit ratio: 1.47 (max 6.36).
- Macro and bank controls:
  - Country controls: lagged real GDP growth, change in interest rate, exchange rate arrangement measure (0 to 6).
  - Bank controls: lagged leverage, lagged loan-to-deposit ratio.
  - Fixed effects: year and country fixed effects.
- Econometric approach:
  - Generalized Method of Moments (GMM) panel regressions with lagged dependent variables (up to lag 4) and instruments to address endogeneity.
  - Policy variables entered as dummies (MaPPj,c,t) for groups and also individually; interactions with lagged ∆Y (policy × ∆Y_i,c,t-1) to capture intensity effects.
  - Phase analysis: upswing vs downswing determined by country real credit increase/decrease in the year; regressions include interaction of policy dummies with downswing dummy to measure differential effects.

### Regression results — main findings
- Sample sizes in regressions after lags/outlier drops:
  - Base regressions: about 2,630 banks and 13,800 observations.
  - Subsequent regressions: about 1,670 banks and 8,500 observations.
- Controls:
  - Lagged dependent variable: negative and statistically significant (convergence/backlash in asset growth).
  - Lagged real GDP growth: positive and statistically significant.
  - Change in interest rate: not consistently significant in base regressions; negative and significant during downswing regressions (lowering rates helps mitigate asset declines).
  - Exchange rate regime: not significant, but more freely-floating regimes have negative sign for asset growth.
  - Lagged leverage: negative and generally statistically significant (higher leverage associated with lower subsequent asset growth).
  - Loan-to-deposit ratio: generally insignificant.
- Policy group effects (Table 5, column 2 — one group at a time):
  - Borrower-based measures (LTV and DTI): reduce asset growth by about 0.44 percentage points (statistically significant).
  - Financial institutions’ asset/liability measures (CG, FC, RR): reduce asset growth by 0.66 percentage points (statistically significant).
  - Buffer-oriented measures (DP, CTC, PRD): no statistically significant effect on asset growth.
  - Other category: negative effect of about 0.67 percentage points (statistically significant).
- Intensity of cycle (column 3): overall results confirmed; no specific statistically significant interactions between policy groups and cycle intensity.
- All four groups simultaneously (column 4): balance-sheets oriented and Other categories show largest impacts, of 1 and 0.72 percentage points respectively.
- Individual-policy regressions (column 5):
  - LTV: statistically significant negative effect of 0.85 percentage points.
  - CG limit: statistically significant negative effect of 0.70 percentage points.
  - CTC: countercyclical capital requirement statistically significant negative.
  - PRD: positive but not statistically significant (appears counterproductive in that regression).
- All nine policies simultaneously (column 6): CG and Other remain important; FC becomes statistically significant negative; LTV loses significance; DTI becomes positive and statistically significant.
- Phase-specific effects (Table 6):
  - Policies are generally more effective in upswings (booms) than in downswings (busts).
  - Borrower-based, balance-sheet-based, and Other policies show significant negative effects on asset growth during booms.
  - In contractionary periods, only borrower-based policies statistically help limit asset growth declines (positive coefficient), suggesting they mitigate declines; other groups generally not significant in downswings.
  - Buffer-building policies do not show significant effectiveness in mitigating downturns (caveat: few countries used these policies).
  - Some balance-sheet-based measures may act perversely during downswings and exacerbate asset declines if not adjusted timely.
- Advanced vs emerging markets (Table 7):
  - Borrower-based measures’ negative effect in full sample is driven more by advanced countries; an emerging-market interaction offsets the negative sign (i.e., borrower-based measures work better in advanced countries).
  - Category Other appears to work well in emerging markets (interaction significant negative).
  - When interacting policies with cycle intensity:
    - Borrower-based policies work mostly for advanced countries, but tend to work more in emerging markets when cycles are intense.
    - Buffer-based policies work more when the cycle is intense in advanced economies, not in emerging markets.
    - Other category works well across countries and irrespective of cycle intensity.
- Overall empirical conclusion:
  - Caps on borrowers (LTV, sometimes DTI) and financial institutions’ asset/liability-based measures (CG, FC, RR) are effective in reducing bank asset growth and thus mitigating vulnerabilities, especially during upswings.
  - Buffer-based instruments (DP, CTC, PRD) show little aggregate impact on asset growth in this sample, with limited evidence they cushion downturns.
  - Little evidence that policy effectiveness varies systematically with cycle intensity, except in some subgroup interactions.

### Policy implications and recommendations (based on findings)
- Macro-prudential policies can be effective elements of the toolkit to reduce banking system vulnerabilities, particularly:
  - Borrower-based measures (LTV, DTI) for dampening excessive lending growth and leverage in booms.
  - Balance-sheet oriented measures (limits on credit growth, foreign currency lending, reserve requirements) to constrain asset growth.
- Policy design should be country- and circumstance-specific:
  - Financial structure (bank- vs market-based), state-owned bank presence, exchange rate regime, and degree of financial openness should inform tool choice.
  - Advanced economies may derive more benefit from borrower-based tools (reflecting real estate driven cycles).
  - Emerging markets may benefit from packages of policies (combinations) and from coordination with capital flow management tools.
- Consideration of potential drawbacks:
  - Circumvention and risk transfer to less-regulated sectors (shadow banking) can offset bank-sector gains — borrower-targeted measures are less easily avoided than bank-targeted ones.
  - Some policies can act perversely in downswings if not timely adjusted (risk of aggravating credit contractions).
  - Adoption of macro-prudential policies may entail resource allocation effects that could affect economic activity and financial sector development.
- Institutional and data needs:
  - Stress tests can complement macro-prudential tools by being forward-looking and more granular.
  - Improved data, supervisory capacity, and political/institutional frameworks are important for effective policy implementation.

### Limitations and suggestions for further research
- Econometric and identification caveats:
  - Residual selection, endogeneity, and omitted variables may remain; suggested methods include propensity scoring, matching banks across regimes, and studying subsidiaries of the same foreign bank operating under different regimes.
- Circumvention and system-wide effects:
  - Need to study risk migration to shadow banking and overall systemic risk measures (aggregate credit, asset prices, systemic risk rankings such as Marginal Expected Shortfall or CoVaR).
- Additional empirical extensions:
  - Use of more aggregate or market-based measures (credit spreads, systemic risk rankings) to capture comprehensive effects and circumvention.
  - Longer time-series and more detailed intra-year policy change data (data constraints limited intra-year analysis here).
- Broader assessment:
  - Further study needed on macroeconomic costs/benefits of macro-prudential policies, including impacts on resource allocation, growth, and financial development.

*Italic: Source — IMF working paper section (analysis and results on effectiveness of macro-prudential policies, sample 2000-2010; bank-level GMM regressions using Bankscope; summary of Tables 2–7 and accompanying discussion).*

### REFERENCES

### _wp14155 - REFERENCES

### Bibliographic references (select highlights)
- Acharya, Viral, 2013, “Adapting Micro-prudential Regulation for Emerging Markets”, Report commissioned by the World Bank’s Poverty Reduction and Economic Management (PREM) Network, May. In Otaviano Canuto and Swati R. Ghosh (eds.), Dealing with the Challenges of Macro Financial Linkages in Emerging Markets, World Bank, Washington, D.C. pp. 57-89.
- Acharya, Viral, Christian Brownlees, Robert Engle, Farhang Farazmand and Mathew Richardson, 2010, “Measuring Sytemic Risk” in Acharya, Viral, Thomas Cooley, and Mathew Richardson (Eds.), Regulating Wall Street: The Dodd-Frank Act and the New Architecture of Global Finance, John Wiley and Sons.
- Adrian, Tobias, and Markus K. Brunnermeier, 2011, “CoVaR,” NBER Working Paper 17454.
- Adrian, Tobias and Hyun S. Shin, 2010, “Liquidity and Leverage,” Journal of Financial Intermediation 19(3), pp. 418-437; 2014, “Procyclical Leverage and Value-at-Risk”, Review of Financial Studies, 27 (2), 373-403.
- Borio, Claudio E.V. and William R. White, 2003, “Whither Monetary and Financial Stability: the Implications of Evolving Policy Regimes?” – in Monetary Policy and Uncertainty: Adapting to a Changing Economy, Conference Volume, Federal Reserve Bank of Kansas City.
- Bank of England, 2009, “The Role of Macroprudential Policy”. Discussion Paper. November; Bank of England, 2011, “Instruments of Macroprudential Policy,” Discussion Paper. December.
- Claessens, Stijn, 2015, “An Overview of Macroprudential Policy Tools”, Annual Review of Financial Economics, forthcoming.
- International Monetary Fund, 2011, “Macroprudential Policy: An Organizing Framework,” Board Paper, April; 2012a, “The Liberalization and Management of Capital Flows — An Institutional View.” IMF Policy Paper, April; 2012b, “The Interaction of Monetary and Macroprudential Policies.” IMF Board Paper and Background Paper, January; 2013a, “Key Aspects of Macroprudential Policy,” IMF Policy Paper, June; 2013b, “Key Aspects of Macroprudential Policy - Background Paper,” IMF Policy Paper, June.
- Lim, C.F. Columba, A. Costa, P. Kongsamut, A. Otani, M. Saiyid, T. Wezel, and X. Wu, 2011, “Macroprudential Policy: What Instruments and How to Use Them? Lessons from Country Experiences,” IMF Working Paper 11/238.

### The Macro-Prudential Toolkit (Table 1: structure and instruments)
- Major policy objectives listed:
  - Enhancing resilience
  - Dampening the cycle
  - Dispelling gestation of cycle
- Instrument groups and example applications in different phases:
  - Restrictions related to borrower, instrument, or activity
    - Time varying caps/limits/rules on: DTI, LTI, LTV; margins, hair-cuts; lending to sectors; credit growth
    - Purpose in expansionary phase: dampening credit; in contractionary phase: adjustment to provisioning, margins; for contagion: varying restrictions on asset composition, activities (e.g., Volcker, Vickers)
  - Restrictions on financial sector balance sheet (assets, liabilities)
    - Time varying caps/limits on mismatches (FX, interest rate); reserve requirements; liquidity limits (e.g., Net Stable Funding Ratio, Liquidity Coverage Ratio)
  - Buffer-based policies
    - Countercyclical capital requirements, leverage restrictions, general (dynamic) provisioning; capital surcharges linked to systemic risk
  - Taxation, levies; Other (including institutional infrastructure)
    - Levy/tax on specific assets and/or liabilities; institutional infrastructure (e.g., CCPs); resolution (e.g., living wills); disclosure and information requirements

### Use of Macro‑Prudential Instruments by Country and Period (Table 2: classification and samples)
- Classification rule:
  - Emerging versus advanced economy countries (source: IMF).
  - Open versus closed capital account countries (source: Chinn-Ito Index 2008). A country is defined as an open capital account country if its Chinn-Ito index is larger than the global median in 2005, and a closed capital account country if its Chinn-Ito index is smaller than the global median in 2005.
- Examples of country-period entries (as listed):
  - Brazil — Closed — Emerging — 2000-2010
  - Canada — Open — Advanced — 2000-2010
  - China — Closed — Emerging — 2000-2010
  - South Korea — Closed — Emerging — 2002-2010
  - Romania — Open — Emerging — 2004-2007
  - Hungary — Open — Emerging — 2010
- Instrument-specific notes (from Table 2 continuation):
  - Foreign currency lending limits (FC): Reduces vulnerability to fx risks; Reduces credit growth directly.
  - Loan-to-value caps (LTV): Reduces vulnerability arising from highly geared borrowings.
  - Debt-to-income caps (DTI): Reduces vulnerability arising from highly geared borrowings.
  - Credit growth caps (CG): Reduces credit growth directly.
  - Reserve requirements (RR): Reduces vulnerability to funding risks; Reduces credit growth indirectly.
  - Profit distribution restrictions (PDR) and other measures (countercyclical provisioning, countercyclical capital): Limit dividend payments in good times to help build up capital buffers in bad times.

### Overall use statistics (Table 3: frequencies and classification)
- There are in total 35 countries using a macro-prudential policy at any point during the period 2000-2010.
- Table 3 frequency-of-use highlights (preserve exact numbers as shown):
  - Loan-to-Value Cap (LTV)
    - Group: 1
    - Total Countries: 24
    - Frequency of Use: 44%
    - Emerging Markets: 15
    - Advanced Countries: 9
    - Closed Capital Account: 11
    - Open Capital Account: 13
    - Frequency of EMs-year: 35%
    - Frequency of ACs-year: 74%
  - Debt-to-Income Ratio (DTI)
    - Group: 1
    - Total Countries: 7
    - Frequency of Use: 9%
    - Emerging Markets: 5
    - Advanced Countries: 2
    - Closed Capital Account: 4
    - Open Capital Account: 3
    - Frequency of EMs-year: 8%
    - Frequency of ACs-year: 11%
  - Credit Growth Caps (CG)
    - Group: 2
    - Total Countries: 6
    - Frequency of Use: 8%
    - Emerging Markets: 5
    - Advanced Countries: 1
    - Closed Capital Account: 4
    - Open Capital Account: 2
    - Frequency of EMs-year: 10%
    - Frequency of ACs-year: 1%
  - Limits on Foreign Lending (FC)
    - Group: 2
    - Total Countries: 8
    - Frequency of Use: 8%
    - Emerging Markets: 7
    - Advanced Countries: 1
    - Closed Capital Account: 4
    - Open Capital Account: 4
    - Frequency of EMs-year: 10%
    - Frequency of ACs-year: 3%
  - Reserve Requirements (RR)
    - Group: 2
    - Total Countries: 5
    - Frequency of Use: 5%
    - Emerging Markets: 5
    - Advanced Countries: 0
    - Closed Capital Account: 5
    - Open Capital Account: 0
    - Frequency of EMs-year: 7%
    - Frequency of ACs-year: 0%
  - Dynamic Provisioning (DP)
    - Group: 3
    - Total Countries: 9
    - Frequency of Use: 9%
    - Emerging Markets: 8
    - Advanced Countries: 1
    - Closed Capital Account: 5
    - Open Capital Account: 4
    - Frequency of EMs-year: 9%
    - Frequency of ACs-year: 11%
  - Counter-cyclical Requirements (CTC)
    - Group: 3
    - Total Countries: 2
    - Frequency of Use: 1%
    - Emerging Markets: 2
    - Advanced Countries: 0
    - Closed Capital Account: 0
    - Open Capital Account: 2
    - Frequency of EMs-year: 0%
    - Frequency of ACs-year: 0%
  - Profit Redistribution (PR)
    - Group: 3
    - Total Countries: 6
    - Frequency of Use: 6%
    - Emerging Markets: 6
    - Advanced Countries: 0
    - Closed Capital Account: 4
    - Open Capital Account: 2
    - Frequency of EMs-year: 4%
    - Frequency of ACs-year: 0%
  - Other MaPP
    - Group: 4
    - Total Countries: 13
    - Frequency of Use: 12%
    - Emerging Markets: 12
    - Advanced Countries: 1
    - Closed Capital Account: 6
    - Open Capital Account: 7
    - Frequency of EMs-year: 15%
    - Frequency of ACs-year: 1%
- Note on classification in Table 3:
  - A country is defined as an open capital account country if its Chinn-Ito Index is larger than the global mean in 2005, and a closed capital account country if its Chinn-Ito Index is smaller than the global mean in 2005.
- Tabulated summary: Total by classification — 35 (only) countries; 25 (only) emerging; 10 (only) advanced; 15 (only) closed capital account; 20 (only) open capital account.

### Summary statistics of regression variables (Table 4A — selected exact entries)
- Sample dimensions and identifiers:
  - Bank Level: 2821 (unique banks), 2000-2010, 48 (unique) Countries
  - ALL SAMPLE: Obs. Leverage Growth (YoY) (%) = 1808; Mean = 20.26; Std. Dev. = 42.88; Min = -385.62; Max = 385.62
  - ALL SAMPLE: Obs. Asset Growth (YoY) (%) = 1809; Mean = 13.32; Std. Dev. = 28.40; Min = -37.55; Max = 77.05
  - Bank-level: Leverage Ratio — Obs. 2122; Mean = 13.55; Std. Dev. = 10.55; Min = 0.84; Max = 39.70
  - Bank-level: Loan to Deposit Ratio — Obs. 1816; Mean = 1.47; Std. Dev. = 1.56; Min = 0.15; Max = 6.36
  - Bank-level: Assets (USD $M) — Obs. 2123; Mean = 458030.42; Std. Dev. = 1151717.10; Min = 0.01; Max = 3914824
  - Country Level: Real GDP Growth (%) — Obs. 2821; Mean = 3.08; Std. Dev. = 3.10; Min = -3.31; Max = 9.52
  - Country Level: Interest Rate Change — Obs. 2277; Mean = 5.34; Std. Dev. = 17.42; Min = -25.80; Max = 43.51
- MaPP subgroup prevalences (Obs. = 310 for many indicators):
  - MaPP subgroup1: Mean = 0.42; Std. Dev. = 0.49
  - LTV: Mean = 0.36; Std. Dev. = 0.43
  - DYI (DTI): Mean = 0.03; Std. Dev. = 0.12
  - CG: Mean = 0.08; Std. Dev. = 0.27
  - FC: Mean = 0.08; Std. Dev. = 0.27
  - RR: Mean = 0.06; Std. Dev. = 0.23
  - DP: Mean = 0.08; Std. Dev. = 0.27
  - CTC: Mean = 0.02; Std. Dev. = 0.15
  - PRD: Mean = 0.02; Std. Dev. = 0.13
  - Other: Mean = 0.09; Std. Dev. = 0.28

### Banking variables by country (Table 4B: selected exact country-level statistics)
- Table 4B reports country-level averages for Leverage, Assets, and NCC (Non-core to Core liabilities) with counts:
  - Argentina: Number of Observations 59, Closed Emerging, Use of MaPP 64, Leverage = 0.73, Assets = 9.13, NCC = -5.97
  - Brazil: Number of Observations 100, Closed Emerging, Use of MaPP 1100, Leverage = 3.25, Assets = 22.20, NCC = 0.07
  - China: Number of Observations 100, Closed Emerging, Use of MaPP 1100, Leverage = -0.54, Assets = 25.29, NCC = -2.78
  - United States: Number of Observations 100, Open Advanced, Use of MaPP 1100, Leverage = -2.40, Assets = 10.17, NCC = -2.54
  - United Kingdom: Number of Observations 100, Open Advanced, Use of MaPP 1100, Leverage = 0.83, Assets = 10.04, NCC = 0.04
  - Russia: Number of Observations 100, Closed Emerging, Use of MaPP 1100, Leverage = 4.39, Assets = 26.11, NCC = -0.68
- Averages reported in the table include:
  - Average (35 MaPP countries)
  - Average (25 emerging countries)
  - Average (23 advanced countries)
  - Average (15 closed capital account countries)
  - Average (33 open capital account countries)
  - Average (13 non-MaPP countries)
  - (Exact numeric summary values are presented in the table columns for each country as shown above.)

### Base regression results on effects of Macro‑Prudential Policies (Table 5: exact coefficient highlights)
- Dependent variable: bank total asset growth; regressions are GMM with instruments described in notes.
- Key estimated coefficients (selected, preserving exact values and significance markers):
  - Lagged Asset Growth:
    - Column (1): -0.346*** [0.051]
    - Column (2): -0.322*** [0.106]
    - Column (3): -0.468** [0.221]
  - Lagged Real GDP Growth (%):
    - Column (1): 4.106*** [1.033]
    - Column (2): 7.705*** [2.829]
    - Column (3): 7.319 [5.022]
  - Lagged Interest Rate Change:
    - Column (1): 0.207** [0.090]
    - Column (2): 0.186 [0.123]
    - Column (3): 0.019 [0.229]
  - Lagged Leverage Ratio:
    - Column (1): -0.022* [0.012]
    - Column (2): -0.029** [0.015]
    - Column (3): -0.024 [0.020]
  - Subgroup2:
    - Column (1): -0.656*** [0.183]
    - Column (2): -0.631*** [0.191]
    - Column (3): -0.994*** [0.344]
  - LTV:
    - Column (3): -0.852*** [0.225]
  - DTI:
    - Column (3): 1.297 [1.126]
    - Column (5): 6.009* [3.128]
  - CG:
    - Column (3): -0.704** [0.346]
    - Column (5): -2.656* [1.532]
  - FC:
    - Column (3): -0.392 [0.244]
    - Column (5): -2.335*** [0.857]
  - CTC:
    - Column (3): -0.406*** [0.099]
    - Column (5): 0.585 [0.521]
  - Other:
    - Column (3): -0.673*** [0.105]
    - Column (5): -1.443** [0.592]
- Observations and panel size:
  - Observations: 13,804 in column (1); 8,527 in columns (2)-(6) (varies by specification).
  - Number of indices (banks): 2,630 in column (1); 1,667 in other columns.
- Note: ***, **, and * denote significance at the 1, 5, and 10 percent levels respectively.

### Effectiveness by phase of the cycle (Table 6: model setup note)
- Dependent variable: bank total asset growth.
- Controls: first lag of asset growth (dependent variable), lagged bank leverage and loan-to-deposit ratios.
- MaPP categories used:
  1. MaPP Aimed at Borrowers (caps on loan-to-value and caps on debt-to-income).
  2. MaPP Aimed at Financial Institutions, Asset Side (limits on credit growth, limits on foreign lending), and Liabilities Side (reserve requirements).
  3. MaPP Aimed at Financial Institutions as Buffers (dynamic provisioning, countercyclical provisioning and countercyclical capital, restrictions on profit distribution).
  4. Other.
- Regressions control for individual trends (country-fixed effects) and use GMM instrumentation (4 lagged differences, lagged real credit growth, and a time trend).

*Source: _wp14155 - REFERENCES*

### 3. MaPP Aimed at Financial Institutions as Buffers (dynamic provisioning, countercyclical provisioning and

### _wp14155 - 3. MaPP Aimed at Financial Institutions as Buffers (dynamic provisioning, countercyclical provisioning and

### Regression results (MaPP Aimed at Financial Institutions as Buffers — GMM estimates)
- Regression setup and significance:
  - These are all GMM regressions which use (4) lagged differences, the lagged real credit growth, and a time trend (fixed effect) as instrumental variables. The regressions control for individual trends (country-fixed effects). GMM standard errors are in brackets. ***, **, and * represent significance at the 1, 5, and 10 percent levels respectively.
- Coefficients (columns correspond to the four regressions shown):
  - Lagged Asset Growth: -0.350*** [,0.072], -0.211** [,0.092], -0.336*** [,0.071], -0.437*** [,0.074]
  - Lagged Real GDP Growth (%): 0.0394 [,1.713], 4.125*** [,1.194], 3.247*** [,1.188], 2.469** [,1.008]
  - Lagged Interest Rate Change: -0.254* [,0.140], 0.061 [,0.212], -0.289* [,0.167], -0.074 [,0.149]
  - Exchange Rate Classification: -3.295** [,1.506], -3.273** [,1.471], -3.232** [,1.488], -2.482*** [,0.863]
  - Lagged Leverage Ratio: 0.015 [,0.013], -0.010 [,0.015], 0.013 [,0.012], -0.010 [,0.012]
  - Lagged Loan to Deposit: -0.072 [,0.114], -0.069 [,0.117], 0.004 [,0.112], 0.056 [,0.104]
  - Downswing: -0.230*** [,0.059], -0.141*** [,0.054], -0.203*** [,0.049], -0.122** [,0.049]
  - Subgroup1: -0.353* [,0.200]
  - Sub1 X downswing: 0.237* [,0.121]
  - Subgroup2: -0.609*** [,0.179]
  - Sub2 X downswing: 0.103 [,0.089]
  - Subgroup3: 0.059 [,0.078]
  - Sub3 X downswing: 0.021 [,0.127]
  - Subgroup4: -0.537*** [,0.100]
  - Sub4 X downswing: 0.188 [,0.134]
- Sample size and panels:
  - Observations: 8,290; 8,290; 8,290; 8,290; 290 (as reported in table)
  - Number of index_number: 1,637; 1,637; 1,637; 1,637; 637
- Note: Standard errors in brackets.

### Table 7: Emerging Markets vs. Advanced Countries (bank total asset growth — GMM estimates)
- Notes on specification:
  - The dependent variable is bank total asset growth. Controls include the first lag of asset growth (the dependent variable), and the lagged bank leverage and loan-to-deposit ratios. The macroprudential policy measures used are: 1. MaPP Aimed at Borrowers (caps on loan-to-value and caps on debt-to-income), 2. MaPP Aimed at Financial Institutions, Asset Side (limits on credit growth, limits on foreign lending), and Liabilities Side (reserve requirements); 3. MaPP Aimed at Financial Institutions as Buffers (dynamic provisioning, countercyclical provisioning and countercyclical capital, restrictions on profit distribution), and 4. Other. Although regressed one at a time, the four MaPPs are shown at the same time in one column in (1) and (3) to save space. The regression in (2) includes all MaPP variables simultaneously. These are all GMM regressions which use (4) lagged differences, the lagged real credit growth, and a time trend (fixed effect) as instrumental variables. The regressions control for individual trends (country-fixed effects). GMM standard errors are in brackets. ***, **, and * represent significance at the 1, 5, and 10 percent levels respectively.
- Coefficients (as reported across columns/panels):
  - Lagged Asset Growth: -0.307** [,0.127]
  - Lagged Real GDP Growth (%): 15.989*** [,5.156]
  - Lagged Interest Rate Change: 0.282* [,0.159]
  - Exchange Rate Classification: -0.545 [,1.225]
  - Lagged Leverage Ratio: -0.008 [,0.019]
  - Lagged Loan to Deposit: 0.311 [,0.197]
  - Subgroup1: -1.274*** [,0.379], -2.473 [,1.508], -1.762*** [,0.518]
  - Sub1 X Emerging: 1.286** [,0.501], 3.622** [,1.690], 1.657*** [,0.634]
  - Sub1 X Lagged Asset Growth: 1.175** [,0.549]
  - Sub1 X Lagged Asset Growth * Emerging: -2.178** [,0.912]
  - Subgroup2: 0.466 [,1.499], 1.201 [,2.487], -5.257 [,3.610]
  - Sub2 X Emerging: -1.226 [,1.589], -2.359 [,2.499], 4.293 [,3.650]
  - Sub2 X Lagged Asset Growth: 17.629 [,12.305]
  - Sub2 X Lagged Asset Growth * Emerging: -17.388 [,12.304]
  - Subgroup3: 0.099 [,0.544], 0.001 [,0.293], 0.457 [,0.798]
  - Sub3 X Emerging: -0.058 [,0.540], -0.500 [,0.798]
  - Sub3 X Lagged Asset Growth: -2.370* [,1.374]
  - Sub3 X Lagged Asset Growth Emerging: 2.196* [,1.283]
  - Subgroup4: -1.008*** [,0.337], 1.306 [,1.213], -0.950** [,0.431]
  - Sub4 X Emerging: 0.380 [,0.355], -2.018* [,1.193], 0.291 [,0.421]
  - Sub4 X Lagged Asset Growth: -0.490 [,1.149]
  - Sub4 X Lagged Asset Growth * Emerging: 0.640 [,1.032]
- Sample size and panels:
  - Observations: 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 8,527; 527 (as reflected in table rows)
  - Number of index number: 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 1,667; 667

### Figures referenced
- Figure 1: Channels Through Which Banks can Become Vulnerable (figure title only).
- Figure 2: Use of Macro‑Prudential Policies: Advanced Countries vs. Emerging Markets and Open vs. Closed Capital Account*
  - Note attached to Figure 2: "*Index of MaPP usage in emerging markets (EMs), advanced countries (ACs), open capital account economies (Open) and closed capital account economies (Closed). The index represents the percentage of countries in our sample that have used macro‑prudential policies. Sources: Lim et al. (2011); Fund staff calculations.*"

*Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp14155.pdf*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp14155.pdf_