## _wp08232

## Source details

**Canonical URL:** [_wp08232](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp08232.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp08232.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp08232.pdf.json)

---

### Introduction and motivation
- The 2007/08 crisis highlighted how banking system losses can tighten credit conditions for companies and households and impose costs on real activity.
- Monetary policymakers used monetary policy to help offset adverse effects of financial instability on real activity; FOMC minutes cited: members agreed that a rate cut of 50 basis points was the most prudent course of action.
- The Federal Reserve reduced the target federal funds rate by a total of 325 basis points until it came to rest at 2 percent, having been reduced by 25 basis points at the meeting in June 2008.
- The June meeting minutes noted a dilemma: significant credit losses and balance sheet pressure constraining credit availability while the near term outlook for inflation had deteriorated and risks to underlying inflation pressures had risen.

### Research questions
- The paper focuses on four empirical questions:
  - (i) To what extent do banks that are weakened by loan losses reduce their loan growth?
  - (ii) How is the strength of this effect affected by initial balance sheet strength?
  - (iii) How does the strength of these effects depend on monetary policy? And how do losses, initial capital buffers and monetary policy interact to bring about changes in credit provision?
  - (iv) Do losses lead to stronger reductions in credit when financing conditions are tight and banks might find access to new capital expensive?

### Theory and hypotheses
- Core theoretical insight: the Modigliani-Miller theorem breaks down for banks; bank capital structure matters for bank investments (loan supply).
- Theoretical channels and implications:
  - Losses tighten banks’ balance sheets and can reduce credit supply.
  - Liquidity rationing and funding constraints are exacerbated by losses.
  - The marginal effect of losses on credit provision is stronger when initial capital is thin.
  - Tight monetary policy can increase the threat of a run and exacerbate balance sheet constraints, reducing lending.
  - Tighter financing conditions and adverse aggregate states amplify credit crunch effects.
- Four hypotheses:
  - (H1) Weakened banks provide less credit than stronger banks.
  - (H2) The effect of loan losses on credit is amplified for banks with little initial capital.
  - (H3) The effect of initial capital on lending is more pronounced if monetary conditions are tight; high initial capital and loose monetary policy are substitutes.
  - (H4) The effects are more pronounced when financing conditions are tight and banks lack easy access to funds (e.g., during crises).

### Data and empirical approach
- Sample:
  - More than 600 listed banks from 32 countries over 1993–2000; regressions use seven years (1994–2000).
  - Countries include Argentina, Australia, Austria, Belgium, Brazil, Canada, Chile, Finland, France, Germany, Greece, Hong Kong, Indonesia, Ireland, Israel, Italy, Japan, Korea Republic, Malaysia, the Netherlands, Norway, Poland, Portugal, Singapore, Spain, Sweden, Switzerland, Thailand, Turkey, the United Kingdom, and the United States.
  - Focus on listed banks to reduce institutional heterogeneity; results apply to large listed banks.
- Key variables and definitions:
  - Loss measured by loan loss provisions (flow of new provisions recorded in the income statement).
  - Capital defined as ratio of total equity capital to noncapital liabilities.
  - Nominal GDP growth used since the dependent variable is nominal loan growth.
- Estimation and identification:
  - Benchmark regression: loan growth on Provisions, nominal GDP growth (GDPG), controls Z, and interactions of Provisions with lagged Capital.
  - Functional form (as in source): 1*)log(−)++++=Δ itititjtitit CAPLossLossGDPGZLoansδγβα  (equation (1) in source).
  - Controls include profitability (RoE), level of capital, country dummies and time-fixed effects.
  - Estimator: feasible generalized least squares (FGLS) allowing for first-order autocorrelation within units and correlation and heteroskedasticity across units.
  - Interaction approach follows Bernanke and Lown (1991), Peek and Rosengren (1995), Kashyap and Stein (2000).
- Crisis classification:
  - Caprio and Klingebiel (2003) crises: Argentina (1995), Brazil (1994-1999), Indonesia (1994 and 1997-), Japan (1991-), Korea (1997-), Malaysia (1997-), Thailand (1997), Poland (1990s) and Turkey (1994).

### Sample statistics (selected, as reported)
- Loan growth: Mean 0.095, Std. Dev. 0.189, Min -0.583, Max 0.962, Obs. 4213, No. of banks 735
- RoE: Mean 0.108, Std. Dev. 0.116, Min -0.798, Max 0.554, Obs. 4811, No. of banks 736
- Provisions: Mean 0.009, Std. Dev. 0.014, Min -0.004, Max 0.132, Obs. 4444, No. of banks 676
- Capital: Mean 0.098, Std. Dev. 0.131, Min 0.013, Max 1.929, Obs. 4913, No. of banks 718
- GDP growth: Mean 0.031, Std. Dev. 0.023, Min -0.069, Max 0.095, Obs. 4913, No. of banks 718
- For the whole sample provisions ratio is just under 1 percent (source statement).

### Main empirical findings (H1 and H2)
- Losses reduce loan growth (evidence for H1):
  - Specification (1): Provisions coefficient = -1.84*** (a 1 percentage point increase in a bank’s loan loss provision ratio reduces its loan growth by 1.84 percentage points).
  - Average loan growth is just under 10 per cent.
- Interaction with initial capital (evidence for H2):
  - Specification (2): Provisions coefficient = -2.85***; Lagged Capital*Provisions coefficient = 6.23***; Capital coefficient = 0.005.
  - Economic magnitude reported: mean capital ratio ≈ 10 per cent; std. dev. of capital = 13 percent; increasing capital by one standard deviation (~0.13) reduces the negative effect of provisions on loan growth by 0.81 (=6.32*0.13) percentage points as reported in the source.
- Other controls:
  - RoE positive and significant (e.g., RoE coefficient ≈ 0.25***), indicating more profitable banks extend more credit.
  - Nominal GDP growth positively correlated with loan growth (e.g., GDP coefficient ≈ 0.43*** in specification (1)).

### Monetary policy effects and interactions
- Monetary policy operationalization:
  - Δr+ (tightening) if short-term rates increased over the year; Δr- (loosening) if rates decreased; Δr0 (neutral) if rates unchanged.
- Direct effects of monetary stance:
  - Δr+ coefficient (column (1) Table 4): -0.02*** (tightening associated with reduced loan growth).
  - Δr- coefficient (column (2) Table 4): 0.02*** (loosening associated with increased loan growth).
  - Δr0 generally not significant.
  - The mean increase in rates, conditional on rates increasing, is 0.67 (67 basis points); the mean decrease, conditional on rates decreasing, is -0.76 (-76 basis points).
  - A 100 basis point increase in rates over a year would reduce loan growth by close to 3 percentage points (0.02/0.67 = .03).
- Interaction of provisions with monetary policy:
  - Provisions*Δr+ coefficient: -2.32*** (tightening amplifies negative effect of provisions).
  - Provisions*Δr- coefficient: 1.74*** (loosening mitigates negative effect of provisions).
  - From specifications (4) and (5): a 1 percentage point increase in provisions leads to:
    - a reduction in loan growth of 1.41 percentage points when monetary policy is loosening;
    - a reduction of 3.68 percentage points when monetary policy is tightening.
- Interaction of capital with monetary policy:
  - Bank capital cushions against the effect of both monetary tightening and monetary easing.
  - Economic magnitude: a one standard deviation (13 percent) increase in a bank’s level of capital roughly halves the effect of monetary policy on loan growth, either way (0.19*0.13 = 0.025).
- Three-way interaction (Provisions × Lagged Capital × Δr):
  - Four key findings:
    1. The effect of provisions on credit growth is increased when monetary policy is tightening and attenuated when monetary policy is loosening.
    2. The effect of provisions on loan growth is reduced when the initial level of capital is high.
    3. High capital cushions against the effect of both tightening and loosening monetary policy.
    4. Initial capital is more important for cushioning the effect of provisions on credit supply when monetary policy is tightening and less important when monetary policy is easing.
  - Interpretation: loose monetary policy and high initial capital are substitutes in cushioning the effect of credit losses on loan supply.

### Crisis versus noncrisis episodes
- Basic credit crunch effects hold both in and outside crises:
  - Loan growth depends positively on nominal GDP growth and RoE and negatively on provisions in both subsamples.
  - Sensitivity of loan growth to provisions does not vary significantly in sign between crisis and noncrisis episodes; effect of losses is pervasive.
- Stronger role of initial capital in crises:
  - Lagged Capital*Provisions coefficient larger for crisis countries: an increase in capital of one standard deviation (13 percent) reduces the effect of losses on loan growth by 1.73 (=13.33*0.13) percentage points for the crisis sample, roughly double the noncrisis effect.
  - Difference is statistically significant when interacting variables with the crisis dummy.
- Monetary policy in crises:
  - Coefficients on monetary policy indicators are similar on full sample and subsample excluding crisis countries (e.g., -0.02 and 0.02).
  - For crisis-country sample, estimated coefficients on monetary policy indicators measure roughly twice the size (about -0.04 and 0.05).
  - Magnitudes of interest rate changes are about twice as large for crisis episodes as for noncrisis episodes (see Tables 13 and 14).
  - Three-way interaction with crisis times:
    - Bank crisis time dummy shows loan growth is reduced during crisis times (Bank crisis time = -0.08*** and -0.09*** in Table 9 columns).
    - Lagged Capital * Provisions * Δr coefficients about -6.86 and 6.82 (similar to Table 6).
    - Interaction of three-way term with crisis indicator:
      - Lagged Capital * Provisions * Δr+ * Bank crisis time = 52.67***
      - Lagged Capital * Provisions * Δr- * Bank crisis time = -27.43***
    - Interpretation: substitution between initial capital and monetary policy in cushioning losses on loan growth is stronger during crisis times.

### Robustness checks
- Bank-fixed effects:
  - Replacing country-fixed effects with bank-fixed effects yields qualitatively and quantitatively similar results.
  - Hausman tests: chi-square = 9.74 (p-value = 0.55) for model (2); chi-square = 13.52 (p-value = 0.26) for model (4).
- Endogeneity and controls:
  - Dropping RoE does not change main conclusions (Table 11), though provisions coefficients become slightly larger in absolute terms and capital becomes significant when RoE omitted.
  - Authors retain specifications that explicitly account for profitability.

### Conclusions and policy implications
- Core conclusions:
  - Loan losses reduce banks’ provision of credit; reduction is larger for banks with smaller initial capital buffers.
  - Losses lead to a stronger reduction in credit when monetary policy is tightening than when it is loosening.
  - Initial capital and loosening monetary policy act as substitutes in cushioning shocks to banks’ balance sheets.
  - Initial capital is more important in cushioning the effect of losses on loan growth during crisis times; interaction between losses, initial capital and monetary policy develops greater force in crisis episodes.
  - Impact of losses and their interaction with capital and monetary policy is present both for crisis countries and noncrisis countries.
- Policy implications:
  - Feedback effects from losses to credit provision are pervasive and may amplify macroeconomic fluctuations even outside full-blown banking crises.
  - Central bank policymakers should account for financial-sector effects given typical policy frameworks and forecasting tools lack a financial sector.
  - Costs of addressing financial instability by monetary policy include moral hazard from expectations of bailouts, conflict with inflation objectives, currency depreciation, and potential loss of exchange rate pegs.
  - High capital and accommodative monetary policy being substitutes implies a moral hazard incentive for banks to rely on ex post monetary support rather than costly capital buffers.
  - Need for prudential measures to increase capitalization in a countercyclical fashion so monetary policy is less constrained ex post.
  - Findings support coordination between monetary policy authorities and supervisory authorities in capital requirement design.

*Source: _wp08232 (IMF PDF).*

### 1. Summary Statistics ..................................................................................................

### _wp08232 - 1. Summary Statistics ..................................................................................................

### Introduction and motivation
- The 2007/08 crisis highlighted how banking system losses can tighten credit conditions for companies and households and impose costs on real activity.
- Monetary policymakers used monetary policy to help offset adverse effects of financial instability on real activity.
- FOMC minutes cited: members agreed that a rate cut of 50 basis points was the most prudent course of action.
- The Federal Reserve reduced the target federal funds rate by a total of 325 basis points until it came to rest at 2 percent, having been reduced by 25 basis points at the meeting in June 2008.
- The June meeting minutes noted a dilemma: significant credit losses and balance sheet pressure constraining credit availability while the near term outlook for inflation had deteriorated and risks to underlying inflation pressures had risen.

### Research questions (explicitly stated)
- The paper focuses on four empirical questions:
  - (i) To what extent do banks that are weakened by loan losses reduce their loan growth?
  - (ii) How is the strength of this effect affected by initial balance sheet strength?
  - (iii) How does the strength of these effects depend on monetary policy? And how do losses, initial capital buffers and monetary policy interact to bring about changes in credit provision?
  - (iv) Do losses lead to stronger reductions in credit when financing conditions are tight and banks might find access to new capital expensive?

### Literature context and contribution
- Builds on literature on credit crunches (e.g., Peek and Rosengren (1995); Cosimano, Chami and Barajas (2004)) and the “bank capital channel” of monetary policy (e.g., Romer and Romer (1990); Kashyap and Stein (2000); Kishan and Opiela (2000); Gambacorta and Mistrulli (2004); Ehrmann et al (2001)).
- Distinctive approach: treats losses (shocks to bank capital) as a key determinant of bank lending behavior, enabling study of interplay between shocks to capital and initial capital levels and the role of monetary policy.
- Closes gap between studies that examine losses but not monetary policy and studies that examine monetary policy and capitalization but not balance sheet shocks.

### Data and empirical approach (summary)
- Uses a unique sample of more than 600 listed banks from 32 different countries over the period 1992- 2000.
- Cross-country design provides counterfactual variation in monetary stance across countries and over time for a given amount of losses at individual banks, helping to alleviate concerns about endogeneity of monetary policy path.

### Main findings
- Controlling for general economic conditions, banks weakened by loan losses extend less credit.
- The size of the reduction in lending depends on banks’ initial balance sheet strength: reductions are weaker for banks with ample capital buffers and stronger for banks with smaller initial capital buffers.
- Losses lead to a stronger reduction in credit when monetary policy is tight.
- In the face of credit losses, tight monetary policy leads to further reductions in credit supply particularly when initial capital is thin.
- Initial capital and loosening monetary policy act as substitutes in cushioning shocks to banks’ balance sheets.
- Initial capital is more important in cushioning the effect of losses on loan growth during crisis times, when financing conditions are tight.
- The interaction between losses, initial capital and monetary policy is stronger in crisis episodes.
- The impact of losses on loan growth and its interaction with initial capital and monetary policy is present both for crisis countries and for countries that did not experience a banking crisis in the period under study.
- Evidence suggests feedback effects from losses to credit conditions are pervasive and that monetary policy affects this feedback both in crisis times and outside crises.
- The results imply credit supply effects may amplify macroeconomic fluctuations more broadly, even outside full-blown banking crises.

### Remarks on applicability to the 2007/08 crisis
- The authors note the 2007/08 episode may be special given the role of securitization: reduced constraints in originate-and-distribute models and later reintermediation pressures on capital and funding.
- Because the analysis relies on balance sheet information and disregards off-balance sheet activity, techniques employed are not easily applied directly to data from the run-up to and unfolding of the 2007/08 crisis.

*Source: _wp08232 - 1. Summary Statistics (IMF PDF).*

### section III. Sections IV-VI present our empirical results in a number of steps. In section IV we

### _wp08232 - section III. Sections IV-VI present our empirical results in a number of steps. In section IV we

### Theory and Hypothesis Development
- Core theoretical insight: the Modigliani-Miller theorem breaks down for banks; bank capital structure matters for bank investments (loan supply).
- Theoretical channels and implications:
  - Losses tighten banks’ balance sheets and can reduce credit supply (Diamond and Rajan (2006); Brunnermeier and Pedersen (2007); Chiesa (2001)).
  - Liquidity rationing and funding constraints are exacerbated by losses (Brunnermeier and Pedersen (2007); Chiesa (2001)).
  - Interaction with initial capital: the marginal effect of losses on credit provision is stronger when initial capital is thin (Chiesa (2001); Brunnermeier and Pedersen (2007)).
  - Monetary policy amplification: tight monetary policy can increase the threat of a run and exacerbate balance sheet constraints, reducing lending (Diamond and Rajan (2006); van den Heuvel (2001); Chiesa (2001)).
  - Financing conditions and aggregate state: tighter financing conditions and adverse aggregate economic conditions amplify credit crunch effects (Stein (1998); Eisfeldt (2004); Chen (1999)).
- Four hypotheses derived from theory:
  - (H1) Provision of credit by banks that are weakened by losses is lower than that of stronger banks.
  - (H2) The effect of loan losses on banks’ provision of credit is amplified for banks with little initial capital.
  - (H3) In the face of losses, the effect of initial capital on lending is more pronounced if monetary conditions are tight. High initial capital and loose monetary policy are substitutes in their effect on bank lending.
  - (H4) In the face of loan losses, the effect of initial capital and monetary policy should be more pronounced if financing conditions more generally are tight and banks do not have easy access to additional funds, such as during crises.

### Empirical Method and Data
- Estimation strategy:
  - Benchmark specification regresses bank loan growth on loan loss provision ratio (Provisions), nominal GDP growth (GDPG), and controls Z; interacts Provisions with lagged Capital.
  - Basic functional form (as in the source):
    - 1*)log(−)++++=Δ itititjtitit CAPLossLossGDPGZLoansδγβα  (equation (1) in source)
  - Loss measured by loan loss provisions (flow of new provisions recorded in income statement).
  - Nominal GDP growth used because dependent variable is nominal loan growth.
  - Controls include bank profitability (RoE), level of capital, full set of country dummies and time-fixed effects.
  - Regressions performed using a feasible generalized least squares (FGLS) estimator allowing for first-order autocorrelation within units and correlation and heteroskedasticity across units.
- Identification considerations:
  - Interaction approach follows Bernanke and Lown (1991), Peek and Rosengren (1995), Kashyap and Stein (2000).
  - GDP growth included to proxy aggregate loan demand and mitigate demand-supply confounding.
  - Interaction of Provisions with lagged Capital used to isolate supply effects (following Peek and Rosengren (1995)).
  - Additional interaction terms introduced to test H3 and H4 (monetary and financial conditions).
- Sample:
  - More than 600 listed banks from 32 countries from 1993 to 2000.
  - Regressions use seven years (1994–2000) because loan growth is dependent variable.
  - Countries listed in the sample include Argentina, Australia, Austria, Belgium, Brazil, Canada, Chile, Finland, France, Germany, Greece, Hong Kong, Indonesia, Ireland, Israel, Italy, Japan, Korea Republic, Malaysia, the Netherlands, Norway, Poland, Portugal, Singapore, Spain, Sweden, Switzerland, Thailand, Turkey, the United Kingdom, and the United States.
  - Focus on listed banks to reduce institutional heterogeneity; results apply to large listed banks (may not generalize to smaller banks and savings institutions).
- Variable definitions and sample statistics (Table 1 summary):
  - Loan growth: Mean 0.095, Std. Dev. 0.189, Min -0.583, Max 0.962, Obs. 4213, No. of banks 735
  - RoE: Mean 0.108, Std. Dev. 0.116, Min -0.798, Max 0.554, Obs. 4811, No. of banks 736
  - Provisions: Mean 0.009, Std. Dev. 0.014, Min -0.004, Max 0.132, Obs. 4444, No. of banks 676
  - Capital: Mean 0.098, Std. Dev. 0.131, Min 0.013, Max 1.929, Obs. 4913, No. of banks 718
  - GDP growth: Mean 0.031, Std. Dev. 0.023, Min -0.069, Max 0.095, Obs. 4913, No. of banks 718
  - Capital defined as ratio of total equity capital to noncapital liabilities (buffer to absorb losses).
- Crises episodes and provisions:
  - Crises countries per Caprio and Klingebiel (2003): Argentina (1995), Brazil (1994-1999), Indonesia (1994 and 1997-), Japan (1991-), Korea (1997-), Malaysia (1997-), Thailand (1997), Poland (1990s) and Turkey (1994).
  - Average provisions across countries and years documented (Table 2 shows country-year provisions up to as high as 0.0855 in Thailand 1999 and other large values).
  - For whole sample provisions ratio is just under 1 percent (source statement).

### Benchmark Results (Tests of H1 and H2)
- Estimation presented in Table 3 (dependent variable: Loan growth):
  - Specification (1):
    - RoE coefficient: 0.25***
    - Provisions coefficient: -1.84***
    - GDP coefficient: 0.43***
    - Year dummies: Yes; Country dummies: Yes
    - No. of obs.: 3460
    - Goodness of fit: 1/ 0.25
  - Specification (2) (adds Lagged Capital*Provisions interaction):
    - RoE coefficient: 0.23***
    - Capital coefficient: 0.005
    - Provisions coefficient: -2.85***
    - Lagged Capital*Provisions coefficient: 6.23***
    - Year dummies: Yes; Country dummies: Yes
    - No. of obs.: 3453
    - Goodness of fit: 0.25
  - Significance notation: *** 1 percent, ** 5 percent, * 10 percent.
- Key empirical findings:
  - Losses reduce loan growth (evidence for H1):
    - A 1 percentage point increase in a bank’s loan loss provision ratio reduces its loan growth by 1.84 percentage points (specification (1)).
    - Average loan growth is just under 10 per cent; an increase in provisions of 1 percentage point is relatively common and amounts to less than one standard deviation (per Table 1).
  - Interaction with initial capital (evidence for H2):
    - Positive Lagged Capital*Provisions coefficient (6.23***) implies that higher initial capital buffers reduce the negative impact of provisions on loan growth; poorly capitalized banks experience a stronger negative effect.
    - Economic magnitude: mean capital ratio ≈ 10 per cent; std. dev. of capital = 13 percent; increasing capital by one standard deviation (~0.13) reduces the negative effect of provisions on loan growth by 0.81 (=6.32*0.13) percentage points as reported in the source.
  - Controls:
    - More profitable banks (RoE) extend more credit (RoE positive and significant).
    - Bank’s level of capital alone not a strong determinant of loan growth in these specifications.
    - Nominal GDP growth positively correlated with loan growth: in specification (1), a 1 percentage point decrease in GDP growth causes loan growth to fall by around 0.43 percentage points.

### The Effect of Monetary Policy (Overview and Initial Findings)
- Monetary policy operationalization:
  - Define Δr+ (tightening) if short-term rates increased over the year; Δr- (loosening) if rates decreased; Δr0 (neutral) if rates unchanged.
  - Use dummy variables Δr+, Δr-, Δr0 to capture policy stance (identify via changes rather than levels).
- Four-step analysis approach:
  1. Check whether monetary policy indicators affect loan growth.
  2. Interact monetary policy indicators with loan losses to see if they modify the effect of losses on loan growth.
  3. Interact monetary policy indicators with level of capital to see if capital conditions determine the effect of monetary policy on loan growth.
  4. Introduce three-way interaction: monetary policy indicator × loan losses × lagged capital.
- Table 4 results (selected coefficients; dependent variable: Loan growth; columns (1)–(6)):
  - RoE: 0.25*** in multiple specifications.
  - Capital: coefficients around 0.006, 0.005, 0.009, 0.004 across specifications.
  - Provisions: coefficients include -1.87***, -1.89***, -1.85***, -1.36***, -3.15***, -1.91*** across columns.
  - GDP: coefficients around 0.45***, 0.42***, 0.43***, 0.46***.
  - Monetary policy (main effects):
    - Δr+ (tightening) coefficient in column (1): -0.02*** (tightening associated with reduced loan growth).
    - Δr- (loosening) coefficient in column (2): 0.02*** (loosening associated with increased loan growth).
    - Δr0 (neutral) generally not significant (no effect).
  - Interactions of Provisions with monetary dummies (selected):
    - Provisions*Δr+ coefficient (in a specification reported): -2.32*** (tightening amplifies negative effect of provisions).
    - Provisions*Δr- coefficient: 1.74*** (loosening mitigates negative effect of provisions).
    - Provisions*Δr0 coefficient: 0.94 (not significant per reporting).
  - Estimation details common across Table 4:
    - Year dummies: Yes; Country dummies: Yes
    - No. of obs.: 3436 in columns reported
    - Goodness of fit: 0.25–0.26 across specifications
    - Significance notation: *** 1 percent, ** 5 percent, * 10 percent.
- Summary interpretation from reported results:
  - Tightening monetary policy is associated with a reduction in loan growth.
  - Loosening monetary policy is associated with an increase in loan growth.
  - Neutral monetary policy does not appear to have a systematic effect on credit growth.
  - Interaction terms indicate monetary stance modifies how provisions affect loan growth: tightening amplifies the negative impact of provisions; loosening mitigates it.

*Source: Authors’ calculations as presented in the supplied content unit.*

### 3. These findings are as one might expect. As regards the size of the effects

### _wp08232 - 3. These findings are as one might expect. As regards the size of the effects

### Effects of monetary policy on loan growth and symmetry
- Effects of tightening and loosening policy are of similar magnitude; tests did not find strong evidence for asymmetry in absolute terms.
- The mean increase in rates, conditional on rates increasing, is 0.67, i.e., 67 basis points over the year.
- The mean decrease in rates, conditional on rates decreasing, is of similar magnitude, at -.76, or -76 basis points over the year.
- A 100 basis point increase in rates over a year would reduce loan growth by close to 3 percentage points (0.02/0.67=.03), with the effect of monetary easing of similar magnitude.
- Results may relate equally to an effect of monetary policy on loan demand and to an effect on loan supply; to investigate loan supply effects requires interactions with bank-level characteristics and arguing independence from loan demand.

### Interaction of provisions (loan losses) with monetary policy
- Weakening balance sheets lead to a stronger reduction in credit growth when monetary policy is tight than when it is loose.
- From specifications (4) and (5): a 1 percentage point increase in provisions leads to:
  - a reduction in loan growth of 1.41 percentage points when monetary policy is loosening;
  - a reduction of 3.68 percentage points when monetary policy is tightening.
- Interpretation: tightening monetary policy exacerbates the effect of weakening bank balance sheets on bank loan supply; easing alleviates it.

### Interaction of bank capital with monetary policy (Table 5 findings)
- Bank capital cushions against the effect of both monetary tightening and monetary easing.
- For banks that are more weakly capitalized, monetary policy has a stronger effect on loan growth.
- Economic magnitude: a one standard deviation (13 percent) increase in a bank’s level of capital roughly halves the effect of monetary policy on loan growth, either way (0.19*0.13=0.025).
- Consistent with the “bank capital channel” literature (e.g., Gambacorta and Mistrulli (2004)): monetary policy may be more powerful when the banking system is weak.

### Three-way interaction: provisions, initial capital, and monetary policy (Table 6 findings)
- Four key findings from Table 6:
  1. The effect of provisions on credit growth is increased when monetary policy is tightening and attenuated when monetary policy is loosening.
  2. The effect of provisions on loan growth is reduced when the initial level of capital is high.
  3. High capital cushions against the effect of both tightening and loosening monetary policy.
  4. The initial level of capital is more important for cushioning the effect of provisions on credit supply when monetary policy is tightening and less important when monetary policy is easing.
- Implication of the fourth finding (in line with hypothesis (H3)): loose monetary policy and high initial capital are substitutes in cushioning the effect of credit losses on loan supply. To offset a given capital shock on credit supply, monetary policy needs to be looser when initial capital is thin, but can stay tighter when initial capital is high.

### Crisis versus noncrisis countries (Section VI)
- During banking crises, banks’ financing conditions tend to be tight; monetary policy shocks may develop greater force in interaction with balance sheet stresses.
- Sample split by Caprio and Klingebiel (2003) crisis-country classification and a time-varying bank crisis time dummy are used to assess differences.
- Basic credit crunch effects (Table 7):
  - Loan growth depends positively on nominal GDP growth and profitability (RoE) and negatively on provisions, both in and outside crises.
  - The sensitivity of loan growth to loan loss provisions does not vary significantly between crisis and noncrisis episodes; effect of losses on loan growth is pervasive.
- Interaction lagged Capital*Provisions:
  - Coefficient larger for crisis countries: an increase in capital of one standard deviation (13 percent) reduces the effect of losses on loan growth by 1.73 (=13.33*0.13) percentage points for crisis sample, roughly double the noncrisis effect.
  - Difference is statistically significant when interacting all variables with the crisis dummy.
  - Interpretation: impact of loan losses on loan growth is stronger for poorly capitalized banks, and more so during a banking crisis.

### Monetary policy effectiveness in crisis vs noncrisis (Table 8 and related tables)
- Coefficients on monetary policy indicators are similar when estimated on the full sample and subsample excluding crisis countries, e.g., (-0.02 and 0.02).
- For crisis-country sample estimated coefficients on monetary policy indicators measure roughly twice the size (at -0.04 and 0.05) compared with noncrisis countries.
- Investigating magnitudes of interest rate changes: magnitude of changes in interest rates are about twice as large for crisis episodes than for noncrisis episodes (see Tables 13 and 14).
- Conclusion: benchmark effect of monetary policy on loan growth is similar for crisis and noncrisis episodes once magnitudes of rate changes are considered.

### Three-way interaction with crisis episodes (Table 9)
- Specification includes bank crisis time dummy interacted with the three-way interaction term (lagged Capital * Provisions * Δr).
- Negative coefficient on crisis indicator: loan growth is reduced during crisis times (Bank crisis time = -0.08*** and -0.09*** in Table 9 columns shown).
- Coefficients on the three-way interaction term (Lagged Capital * Provisions * Δr) are, at -6.86 and 6.82, similar to Table 6.
- Interaction of three-way term with the crisis indicator shows the effect is considerably larger during banking crises:
  - Lagged Capital * Provisions * Δr+ * Bank crisis time = 52.67*** (column shown);
  - Lagged Capital * Provisions * Δr- * Bank crisis time = -27.43*** (column shown).
- Interpretation: initial capital and monetary policy are substitutes in cushioning losses on loan growth, and this substitution effect is stronger during crisis times.

### Robustness checks (Section VII)
- Bank-fixed effects:
  - Replacing country-fixed effects with bank-fixed effects yields qualitatively and quantitatively similar results to earlier tables (e.g., Table 4).
  - Hausman test statistics: chi-square = 9.74 (p-value = 0.55) for model (2); chi-square = 13.52 (p-value = 0.26) for model (4). These indicate inability to reject that coefficients are the same across random and fixed effects; country- and time-fixed effects are generally sufficient.
- Endogeneity of bank-specific characteristics:
  - Dropping RoE does not change main conclusions (Table 11), though coefficients on provisions appear slightly bigger (in absolute terms) and capital becomes significant, suggesting omitted variable bias when RoE omitted.
  - Preference retained for specification that explicitly accounts for profitability.

### Conclusions (Section VIII) — main qualitative findings
- Loan losses result in a reduction of banks’ provision of credit; reduction is larger for banks with smaller initial capital buffers, all else equal.
- Losses lead to a stronger reduction in credit when monetary policy is tightening than when it is loosening. In the face of credit losses, tightening monetary policy leads to further reductions in credit particularly when initial capital is thin. Initial capital and loosening monetary policy act as substitutes in cushioning shocks to banks’ balance sheets.
- Initial capital is more important in cushioning the effect of losses on loan growth during crisis times; interaction between losses, initial capital and monetary policy develops greater force in crisis episodes.
- Impact of losses on loan growth and its interplay with initial capital and monetary policy is present both for crisis countries and for noncrisis countries.

### Policy implications
- Feedback effects from losses to the provision of credit are pervasive and not confined to specific episodes.
- These feedbacks may amplify macroeconomic fluctuations even outside full-blown banking crises; central bank policymakers should account for financial-sector effects given typical policy frameworks and forecasting tools lack a financial sector.
- Costs of addressing financial instability by monetary policy include moral hazard from expectations of bailouts, conflict with inflation objectives, currency depreciation, and potential loss of exchange rate pegs.
- Finding that high capital and accommodative monetary policy are substitutes in relieving impact of losses implies:
  - Moral hazard incentive for banks to rely on ex post monetary support rather than costly capital buffers.
  - Need for prudential measures to increase capitalization in a countercyclical fashion so monetary policy is less constrained ex post.
- Results support the notion that financial stability and monetary policy are intimately related and suggest central banks have stake in the design of capital requirements and coordination with supervisory authorities.

*Source: _wp08232 - 3. These findings are as one might expect. As regards the size of the effects (PDF chapter/section).*

### References

### References

### References list

- Altunbas, Yener, Leonardo Gambacorta, and David Marqués, 2007, “Securitization and the Bank Lending Channel,” ECB Working Paper, No. 838, December (published also as Working Paper of the Bank of Italy, No. 653).  
- Bayoumi, Tamim, and Ola Melander, 2008, “Credit Matters: Empirical Evidence on U.S. Macro-Financial Linkages,” IMF WP/08/169 (Washington: International Monetary Fund).  
- Bernanke, Ben S., and Cara S. Lown, 1991, “The Credit Crunch,” Brookings Papers on Economic Activity, Vol. 2:1991, pp. 205-47.  
- Borio, Claudio, and William White, 2003, “Whither Monetary and Financial Stability? The Implications of Evolving Policy Regimes,” Monetary Policy and Uncertainty: Adopting to a Changing Economy, Jackson Hole Symposium, Federal Reserve Bank of Kansas City, (August).  
- Brunnermeier, Markus K., and Lasse H. Pedersen, 2007, “Market Liquidity and Funding Liquidity,” Review of Financial Studies, (forthcoming).  
- Chava, Sudheer, and Amiyatosh K., Purnanandan, 2005, “The Effect of Banking Crisis on Bank-dependent Borrowers,” mimeo, Bauer School of Business, University of Houston and Ross School of Business, University of Michigan.  
- Caprio, Gerard, and Daniela Klingebiel, 2003, “Episodes of Systemic and Borderline Financial Crises,” manuscript, World Bank.  
- Chen, Yehning, 1999, “Banking Panics: The Role of the First-come, First-served Rule and Information Externalities,” Journal of Political Economy, Vol. 107 (5), pp. 946-968.  
- Chiesa, Gabriella, 2001, “Incentive-based Lending Capacity, Competition and Regulation in Banking,” Journal of Financial Intermediation, Vol. 10 (1),  pp. 28-53.  
- Cosimano, Thomas F., Ralph Chami, and Adolfo Barajas, 2004, “Did the Basel Accord Cause a Credit Slowdown in Latin America?” ECONOMIA: The Journal of the Latin American and Caribbean Economic Association.  
- Diamond, Douglas W., and Raghuram G. Rajan, 2006, “Money in a Theory of Banking,” American Economic Review, Vol. 96 (1), pp. 30-53.  
- Dell’Ariccia, Giovanni, Enrica Detragiache, and Raghuram G. Rajan, 2008, “The Real Effect of Banking Crises,” Journal of Financial Intermediation, Vol. 17 (1), pp. 89-112.  
- Ehrmann, Michael, Leonardo Gambacorta, Jorge Martinez-Pages, Patrick Sevestre, and Andreas Worms, 2001, “Financial Systems and the Role of Banks in the Monetary Policy Transmission in the Euro Area,” ECB Working Paper No. 105.  
- Eisfeldt, Andrea L., 2004, “Endogenous Liquidity in Asset Markets,” Journal of Finance, Vol. 59 (1), pp. 1-30.  
- Estrella, Arturo, 2002, “Securitization and the Efficacy of Monetary Policy,” FRBNY Economic Policy Review, (May).  
- Gambacorta, Leonardo, and Paolo E. Mistrulli, 2004, “Does Bank Capital Affect Lending Behaviour,” Journal of Financial Intermediation, Vol. 13 (4), pp. 436-57.  
- Giannetti, Mariassunta, 2007, “Financial Liberalization and Banking Crises: The Role of Capital Inflows and Lack of Transparency,” Journal of Financial Intermediation, Vol. 16 (1), pp. 32-63.  
- Gorton, Gary, and Andrew Winton, 2002, “Financial Intermediation,” NBER Working Paper 8928. Available at:  http://www.nber.org/papers/w8928.  
- Kashyap, Anil K., and Jeremy C. Stein, 2000, “What do a Million Observations of Banks Say About the Transmission of Monetary Policy?” American Economic Review, Vol. 90 (3), pp. 407-28.  
- Kishan, Ruby P., and Timothy P. Opiela, 2000, “Bank Size, Bank Capital and the Bank Lending Channel,” Journal of Money, Credit and Banking, Vol. 32, pp. 121-41.  
- Kroszner, Randall S., Luc Laeven, and Daniela Klingebiel, 2007, “Banking Crises, Financial Dependence and Growth,” Journal of Financial Economics, Vol. 84 (1), pp. 187-228.  
- Loutskina, Elena, and Philip E. Strahan, 2006, “Securitization and the Declining Impact of Bank Finance on Loan Supply: Evidence from Mortgage Acceptance Rates,” NBER Working Paper, No. 11983.  
- Lown, Cara S., and Donald Morgan, 2006, “The Credit Cycle and the Business Cycle: New Findings Using the Loan Officer Opinion Survey,” Journal of Money, Credit and Banking, Vol. 38 (6), pp. 1575-97.  
- Pain, Darren, 2003, “The Provisioning Experience of the Major UK Banks: A Small Panel Investigation,” Bank of England Working Paper No. 177.  
- Peek, Joe, and Eric S. Rosengren, 1995, “The Capital Crunch: Neither a Borrower nor a Lender Be,” Journal of Money, Credit and Banking, Vol. 27, No. 3, pp. 625-38.  
- Peek, Joe, and Eric S. Rosengren, 2000, “Collateral Damage: Effects of the Japanese Bank Crisis on Real Activity in the United States,” American Economic Review, Vol. 90,  pp. 30-45.  
- Peek, Joe, and Eric S. Rosengren, 2005, “Unnatural Selection: Perverse Incentives and the Misallocation of Capital in Japan,” American Economic Review, Vol. 95, pp. 1144-1166.  
- Peek, Joe, Eric S. Rosengren, and Geoffrey M. B. Tootell, 2003, “Does the Federal Reserve Possess an Exploitable Informational Advantage? Journal of Monetary Economics, Vol. 50, pp. 817-839.  
- Pilotte, Eugene, 1992, “Growth Opportunities and the Stock Price Response to new Financing,” Journal of Business, Vol. 65, pp. 371-394.  
- Romer, Christina D., and David H. Romer, 1990, “New Evidence on the Monetary Transmission Mechanism,” Brookings Papers on Economic Activity 1990 No. 1, pp. 149-98.  
- Stein, Jeremy C., 1998, “An adverse Selection Model of Bank Asset and Liability Management with Implications for the Transmission of Monetary Policy,” RAND Journal of Economics, Vol. 29, pp. 466-86.  
- Van den Heuvel, Skander J, 2001, “The Bank Capital Channel of Monetary Policy,” University of Pennsylvania, mimeo.

*Source: _wp08232 - References (PDF).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp08232.pdf_