## _wp10131 - References

## Source details

**Canonical URL:** [_wp10131 - References](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp10131.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp10131.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp10131.pdf.json)

---

### Introduction and context
- Sample period covered: 2006 Q1 to 2009 Q2.
- Aggregate lending dynamics:
  - Private sector credit from commercial banks slowed from annual rates of "8 percent or greater" (2003 through Q1 2008) to "just over 2 percent" by end–2008, and thereafter registered negative growth.
  - Non–bank institutions reached growth rates of "close to 19 percent" in early 2004 and in recent months were contracting in nominal terms by "over 11 percent".
- Federal Reserve policy actions:
  - Starting in August 2007, the Federal Reserve cut interest rates totaling "225 basis points" over seven months.
  - Policy sequence: (i) liquidity provision to solvent institutions during interbank market contraction (illiquidity spiral), then (ii) direct liquidity to key borrowers and investors when credit risk became primary.
- TARP / CPP capital injections:
  - TARP allocated "$250 billion out of the total $700 billion" to the Capital Purchase Program (CPP).
  - Initial injection: "$125 billion" into nine large institutions on October 14, 2008.
  - CPP broadened to include "over 550 smaller institutions".
  - Overall injection: "just over $200 billion" as of October 2009.
  - Redemption of shares: "$70.4 billion"; CPP funds outstanding: "$134.1 billion" as of mid–October 2009.
- Excess reserves and Fed balance sheet:
  - Excess reserves: about "$1.5 billion" throughout 2007 and most of 2008; rose to "$900 billion" by January 2009; remained above "$800 billion" through September 2009.
  - Federal Reserve balance sheet increased by "$1,142 billion" amid shadow banking interventions.

### Theoretical framing and hypotheses
- Two main theoretical approaches:
  - Peek and Rosengren (1995): bank intermediation framework yielding testable implications that differ when capital is binding vs non–binding.
  - Chami and Cosimano (2001, 2010): capital channel of monetary policy where interest rate shocks transmit to credit supply via regulatory capital constraints; bank capital choice modeled as a call option on future loan supply.
- Key conceptual implications:
  - Capital constraints operate nonlinearly, with greater effects on banks closer to regulatory minimums.
  - Capital constraints affect lending disproportionately relative to other assets (banks may reallocate to low risk–weighted assets even when total scale expands).
  - Monetary policy design must account for regulatory capital constraints as a channel affecting bank lending.

### Data and descriptive statistics
- Data source: Bankscope bank–level quarterly data.
- Focus: largest bank holding companies by total assets; rationale:
  - Largest institutions represent the vast majority of commercial bank assets.
  - Call report data for the 88 commercial banks with over "$10 billion" in assets showed these banks held over "76 percent" of all commercial bank assets and over "73 percent" of loans and deposits during the fourteen quarters ending in 2009 Q2.
- Behavior of size groups:
  - For the 11–100 largest holding companies: visible slowdowns in total assets (TA), loans, and deposit and short term funding (DSTF); TA and DSTF slow more sharply than loans.
  - For the ten largest institutions: TA and DSTF went from growth of "over 20 percent" in early 2006 to an annual decline of "10 percent" by end–2008.
- Bank loans:
  - Aggregate bank loans decelerated by "almost 6 percentage points" during the sample.
  - For the ten largest banks the slowdown was "by ten percentage points".
  - In 2009 the slowdown intensified, with an annual decline of "4 percent" by the second quarter for the total banking system.
- Capital trends:
  - Although regulatory minimum capital requirements were being met on average, there was a noticeable fall in bank capital between 2006 and the last quarter of the sample period.

### Empirical exercises and key findings
- Three empirical exercises:
  1. Peek and Rosengren (1995) test for binding capital constraint using cross–section variability and pre–crisis heterogeneity across banks.
     - Finding: evidence supports that banks were capital constrained from 2007 to 2009—banks with lower initial capital ratios were more likely to contract operations and restrict credit.
     - Robustness: result holds across different capital measures—tier 1, total capital, and the leverage ratio.
     - Liquidity: no evidence linking initial liquidity (conventional measure: initial liquid asset ratio) to subsequent contractions in banking activities; caveat that conventional liquidity measures may understate the true liquidity crunch from interbank market collapse.
  2. Test of the Chami–Cosimano model for bank capital choice.
     - Theoretical prediction: a negative and convex relation between new bank capital and previous changes in bank capital, interest expenses, and non–interest expenses.
     - Empirical result: confirms the model’s predictions.
  3. Test of monopoly power (Panzar and Rosse H–test).
     - Finding: H–test supports monopolistic–competitive banking behavior during the study period.
     - Empirical observation: bank holding companies systematically increased their net interest margin throughout 2007–2008 while the Federal Reserve reduced short–term interest rates by "5 percentage points".
- Main regression results and robustness:
  - Capital as a constraint:
    - Coefficients on lagged change in capital–asset ratio and its interaction with initial capital are positive and negative respectively, and both are statistically significant for both loan and deposit growth.
    - Results robust to measure of capital used (Total capital, Tier 1, or equity–asset ratio).
    - Stronger relationships during the capital loss period (2006 Q4–2008 Q4) than in the extended sample including 2009 H1.
    - Conclusion: capital, rather than liquidity (as measured by liquid asset ratio), was the primary constraint on lending during the crisis period analyzed.
  - Liquidity:
    - Initial liquid asset ratio did not significantly predict higher subsequent deposit or loan growth.
    - Caveat: liquid asset ratio may poorly capture funding stress caused by interbank market disruptions and extensive use of repo funding.
  - TARP effectiveness:
    - Ratio of TARP funds to total assets was not significant in regressions; TARP funds distributed in 2008 Q4 did not have an observable impact on loan supply in the short sample window analyzed (including one–quarter lags).
    - Consistent evidence from other work: of every new dollar of TARP capital injections (CPP flows), only between 10 and 18 cents went toward new lending, whereas about 67 cents went to boosting banks’ Tier 1 capital ratios.
  - Choice of bank capital (CC model tests):
    - Banks increased capital after past capital losses; the effect is larger for banks that were more capital constrained initially.
    - Quarterly bank stock return is positively related to Tier 1 and total capital ratios (sample where available).
    - Interest expense significant; non–interest expense and nonperforming loans show expected sign patterns with varying significance.
    - Time dummies indicate a generalized uptick in regulatory bank capital in 2008 Q4 that persisted into H1 2009 (partly due to TARP and balance sheet risk–weight reductions).
    - Model explains at least 63 percent of variation in total bank capital (specifications excluding quarterly stock return).
  - Monopoly power / H–statistic (Table 6 highlights):
    - OLS coefficients (robust standard errors in parentheses):
      - Interest expense ratio 0.51  * * * (0.07)
      - Interest expense ratio · Capital Constraint Dummy 0.07 (0.119)
      - Noninterest expense ratio 0.09 *** (0.029)
      - Noninterest expense ratio · Capital Constraint Dummy -0.09 (0.082)
      - Ratio of nonperforming loans to assets 0.00 (0.000)
      - Nonperforming loan ratio · Capital Constraint Dummy 0.00 (0.001)
      - Capital Constraint Dummy -0.07 (0.176)
      - Logarithm of assets -0.01 (0.014)
      - Dummy =1 for banks whose assets >$17 billion and <$460 billion -0.02 (0.062)
      - Dummy =1 for banks whose assets <$17 billion 0.06 (0.080)
      - Dummy for 2007Q3 0.23  * * * (0.029)
      - Dummy for 2007Q4 0.18  * * * (0.029)
      - Dummy for 2008Q1 0.12  * * * (0.025)
      - Dummy for 2008Q2 0.08  * * * (0.023)
      - Dummy for 2008Q3 0.07  * * * (0.026)
      - Constant 0.84  * * * (0.138)
      - R-squared 0.5746
    - Tests on the coefficients: F-statistic (p-value)
      - H0: Sum of the coefficients on input prices = 0
        - Capital constraint not binding 45.87 (0.000)
        - Capital constraint binding 12.26 (0.000)
      - H0: Sum of the coefficients on input prices = 1
        - Capital constraint not binding 20.89 (0.000)
        - Capital constraint binding 6.56 (0.010)
    - Interpretation: tests reject H = 0 and H = 1, implying the sum of coefficients lies between zero and one—consistent with monopolistic competition among the largest bank holding companies.

### Capital, liquidity, and regulatory context (2006 Q4–2008 Q4)
- Equity–asset ratio: declined by less than 1 percentage point for the system as a whole during 2006 Q4–2008 Q4.
- Total capital ratio: remained above regulatory minimums for the system as a whole and did not suffer a similar decline across all groups; the largest bank holding companies registered noticeable increases in total capital (partly due to TARP injections and asset recomposition away from loans).
- Liquid asset ratio: rose by 2–3 percentage points during the crisis period discussed.
- Regulatory minimums and classification thresholds:
  - Tier 1 minimum: 4 percent of risk–weighted assets.
  - Total capital minimum: 8 percent of risk–weighted assets.
  - FDIC “well capitalized” thresholds: at least 6 and 10 percent, respectively (Tier 1 and Total Capital).
  - Minimum equity–asset ratio: 3 percent.
- Federal funds rate: lowered by 5 percentage points between August 2007 and December 2008, from 5.25 percent to 0.25 percent.

### Profitability decline and decomposition
- System and 88 largest banks: ROE dropped by about 12 percentage points between 2006 Q4 and 2008 Q4, from about 13 percent to just over 1 percent.
- Decomposition result: most of the decline in profitability can be attributed to a 1–percentage point increase in the loan loss provision ratio amplified by an equity multiplier (A/E) of just over 10.
- Net interest margin (NIM): decreased only slightly, by about 0.2 percentage points; this was almost entirely offset by a similar decline in non–interest expense.
- Off–balance sheet items: significant decline consistent with the fall in fee-collecting activities.
- For the largest ten bank holding companies:
  - ROE fell into negative territory by the end of 2007 and declined by about 43 percentage points between 2006 and 2008.
  - Securities losses accounted for close to 30 percentage points of the ROE decline once amplified by an equity multiplier of about 16.
  - Contraction in NIM contributed about 4 percentage points.
  - Increase in noninterest expenses contributed about 4 percentage points.
  - Increase in provisions for loan losses contributed about 5 percentage points.
- Accounting and measurement caveats:
  - TARP injections are included within total capital, but not counted as equity.
  - Commercial banks are not obliged to mark to market all assets; thus, securities losses may be recognized differently across entities.

### Policy implications and recommendations
- Monetary policy design:
  - Monetary policy must account for regulatory capital constraints because bank capital is a crucial conduit for monetary transmission to bank lending.
  - In environments where banks must hold a percentage of risky assets as capital, banks cannot expand lending without additional capital; they need incentives in the form of higher expected future returns to raise costly capital.
- Unwinding extraordinary measures and privatization of Fed-held assets:
  - The Federal Reserve acquired "$1,142 billion" of additional assets held by the shadow banking system during the crisis.
  - Privatization of these assets should avoid replaying the crisis; assets should be sold to private intermediaries subject to prudential regulation.
  - To privatize without destabilizing banks, "$114.2 billion" in total capital—10 percent of the required balance sheet expansion—must be added to the banking system so that institutions are "well qualified" under current commercial bank regulations.
  - Banks must receive clear signals that they can acquire these assets at a competitive risk–adjusted rate of return to incentivize optimal capital raising without excessive risk-taking.
  - Higher loan rates required to fund these assets would help mitigate excessive borrower risk.

### Overall conclusions
- Evidence from 2006 Q1–2009 Q2 supports the conclusion that capital constraints were a primary factor explaining the collapse in bank lending during the 2007–2009 financial crisis.
- Conventional liquidity measures did not predict bank lending contractions in this sample, though measurement caveats exist.
- The empirical validation of theoretical models implies that policies aimed at restoring bank capital and aligning incentives for capital issuance are fundamental to reactivating credit.

*Source: _wp10131 - References (excerpt) — content from the provided IMF PDF chapter/section.*

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

### _wp10131 - References

### Introduction and context
- Sample period covered: 2006 Q1 to 2009 Q2.
- Aggregate lending dynamics:
  - Private sector credit from commercial banks slowed from annual rates of "8 percent or greater" (2003 through Q1 2008) to "just over 2 percent" by end–2008, and thereafter registered negative growth.
  - Non–bank institutions reached growth rates of "close to 19 percent" in early 2004 and in recent months were contracting in nominal terms by "over 11 percent".
- Federal Reserve policy actions:
  - Starting in August 2007, the Federal Reserve cut interest rates totaling "225 basis points" over seven months.
  - Policy sequence: (i) liquidity provision to solvent institutions during interbank market contraction (illiquidity spiral), then (ii) direct liquidity to key borrowers and investors when credit risk became primary.
- TARP / CPP capital injections:
  - TARP allocated "$250 billion out of the total $700 billion" to the Capital Purchase Program (CPP).
  - Initial injection: "$125 billion" into nine large institutions on October 14, 2008.
  - CPP broadened to include "over 550 smaller institutions".
  - Overall injection: "just over $200 billion" as of October 2009.
  - Redemption of shares: "$70.4 billion"; CPP funds outstanding: "$134.1 billion" as of mid–October 2009.
- Excess reserves and Fed balance sheet:
  - Excess reserves: about "$1.5 billion" throughout 2007 and most of 2008; rose to "$900 billion" by January 2009; remained above "$800 billion" through September 2009.
  - Federal Reserve balance sheet increased by "$1,142 billion" amid shadow banking interventions.

### Theoretical framing and hypotheses
- Two main theoretical approaches:
  - Peek and Rosengren (1995): bank intermediation framework yielding testable implications that differ when capital is binding vs non–binding.
  - Chami and Cosimano (2001, 2010): capital channel of monetary policy where interest rate shocks transmit to credit supply via regulatory capital constraints; bank capital choice modeled as a call option on future loan supply.
- Key conceptual implications:
  - Capital constraints operate nonlinearly, with greater effects on banks closer to regulatory minimums.
  - Capital constraints affect lending disproportionately relative to other assets (banks may reallocate to low risk–weighted assets even when total scale expands).
  - Monetary policy design must account for regulatory capital constraints as a channel affecting bank lending.

### Data and descriptive statistics
- Data source: Bankscope bank–level quarterly data.
- Focus: largest bank holding companies by total assets; rationale:
  - Largest institutions represent the vast majority of commercial bank assets.
  - Call report data for the 88 commercial banks with over "$10 billion" in assets showed these banks held over "76 percent" of all commercial bank assets and over "73 percent" of loans and deposits during the fourteen quarters ending in 2009 Q2.
- Behavior of size groups (total commercial banking system, ten largest bank holding companies, 11th–100th holding companies):
  - For the 11–100 largest holding companies: visible slowdowns in total assets (TA), loans, and deposit and short term funding (DSTF); TA and DSTF slow more sharply than loans.
  - For the ten largest institutions: TA and DSTF went from growth of "over 20 percent" in early 2006 to an annual decline of "10 percent" by end–2008.
- Bank loans:
  - Aggregate bank loans decelerated by "almost 6 percentage points" during the sample.
  - For the ten largest banks the slowdown was "by ten percentage points".
  - In 2009 the slowdown intensified, with an annual decline of "4 percent" by the second quarter for the total banking system.
- Capital trends:
  - Although regulatory minimum capital requirements were being met on average, there was a noticeable fall in bank capital between 2006 and the last quarter of the sample period.

### Empirical exercises and key findings
- Three empirical exercises:
  1. Peek and Rosengren (1995) test for binding capital constraint using cross–section variability and pre–crisis heterogeneity across banks.
     - Finding: evidence supports that banks were capital constrained from 2007 to 2009—banks with lower initial capital ratios were more likely to contract operations and restrict credit.
     - Robustness: result holds across different capital measures—tier 1, total capital, and the leverage ratio.
     - Liquidity: no evidence linking initial liquidity (conventional measure: initial liquid asset ratio) to subsequent contractions in banking activities; caveat that conventional liquidity measures may understate the true liquidity crunch from interbank market collapse.
  2. Test of the Chami–Cosimano model for bank capital choice.
     - Theoretical prediction: a negative and convex relation between new bank capital and previous changes in bank capital, interest expenses, and non–interest expenses.
     - Empirical result: confirms the model’s predictions.
  3. Test of monopoly power (Panzar and Rosse H–test) following Cetorelli and Gambera (2001) and Claessens and Laeven (2004).
     - Finding: H–test supports monopolistic–competitive banking behavior during the study period.
     - Empirical observation: bank holding companies systematically increased their net interest margin throughout 2007–2008 while the Federal Reserve reduced short–term interest rates by "5 percentage points".

### Policy implications and recommendations
- Monetary policy design:
  - Monetary policy must account for regulatory capital constraints because bank capital is a crucial conduit for monetary transmission to bank lending.
  - In environments where banks must hold a percentage of risky assets as capital, banks cannot expand lending without additional capital; they need incentives in the form of higher expected future returns to raise costly capital.
- Unwinding extraordinary measures:
  - As recovery takes hold, attention should be paid to how to unwind extraordinary monetary and balance sheet measures (including Fed holdings of shadow banking assets).
  - The paper suggests ways that excess funds on the Federal Reserve balance sheet might be effectively privatized and sold to banks, recognizing the central role of bank capital in enabling lending.

### Overall conclusions
- Evidence from 2006 Q1–2009 Q2 supports the conclusion that capital constraints were a primary factor explaining the collapse in bank lending during the 2007–2009 financial crisis.
- Conventional liquidity measures did not predict bank lending contractions in this sample, though measurement caveats exist.
- The empirical validation of theoretical models implies that policies aimed at restoring bank capital and aligning incentives for capital issuance are fundamental to reactivating credit.

*Source: _wp10131 - References (excerpt) — content from the provided IMF PDF chapter/section.*

### 2008. The equity asset ratio declined for both commercial banks and for the largest bank holding

### 2008. The equity asset ratio declined for both commercial banks and for the largest bank holding companies. However, the total capital ratio remained above the regulatory minimum and did not suffer a similar decline across all groups.

### Capital, liquidity, and regulatory context
- Equity–asset ratio: declined by less than 1 percentage point for the system as a whole during 2006 Q4–2008 Q4.
- Total capital ratio: remained above regulatory minimums for the system as a whole and did not suffer a similar decline across all groups; the largest bank holding companies registered noticeable increases in total capital (partly due to TARP injections and asset recomposition away from loans).
- Liquid asset ratio: rose by 2–3 percentage points during the crisis period discussed.
- Regulatory minimums and classification thresholds:
  - Tier 1 minimum: 4 percent of risk–weighted assets.
  - Total capital minimum: 8 percent of risk–weighted assets.
  - FDIC “well capitalized” thresholds: at least 6 and 10 percent, respectively (Tier 1 and Total Capital).
  - Minimum equity–asset ratio: 3 percent.
- Federal funds rate: lowered by 5 percentage points between August 2007 and December 2008, from 5.25 percent to 0.25 percent.

### Profitability decline and decomposition (ROE → ROA × equity multiplier)
- System and 88 largest banks: ROE dropped by about 12 percentage points between 2006 Q4 and 2008 Q4, from about 13 percent to just over 1 percent.
- Decomposition result: most of the decline in profitability can be attributed to a 1–percentage point increase in the loan loss provision ratio amplified by an equity multiplier (A/E) of just over 10.
- Net interest margin (NIM): decreased only slightly, by about 0.2 percentage points; this was almost entirely offset by a similar decline in non–interest expense.
- Off–balance sheet items: significant decline consistent with the fall in fee-collecting activities.
- For the largest ten bank holding companies:
  - ROE fell into negative territory by the end of 2007 and declined by about 43 percentage points between 2006 and 2008.
  - Securities losses accounted for close to 30 percentage points of the ROE decline once amplified by an equity multiplier of about 16.
  - Contraction in NIM contributed about 4 percentage points.
  - Increase in noninterest expenses contributed about 4 percentage points.
  - Increase in provisions for loan losses contributed about 5 percentage points.
- Accounting and measurement caveats:
  - TARP injections are included within total capital, but not counted as equity.
  - Commercial banks are not obliged to mark to market all assets; thus, securities losses may be recognized differently across entities.

### Empirical approach: specifications and theoretical bases
- Peek and Rosengren (PR) framework:
  - Tests whether initial capital and changes in capital constrain deposit and loan growth.
  - Key expectation: banks with lower initial capital and larger capital losses should show greater contractions in deposits and loans.
  - Empirical regression form (deposit growth example): ΔA_D/A = a0 + a1 Δ(K/A) + a2 [Δ(K/A) × (K/A)_initial] + a3 (K/A)_initial + controls + ε.
- Chami and Cosimano (CC) model:
  - Treats capital as an option (call option) whose strike price depends on future loan demand shock; banks hold extra capital for flexibility.
  - Predicts convexity: demand for capital increases when initial capital is lower and when interest and non–interest expenses are lower; implies nonlinear responses to prior capital changes.
  - Empirical specification links current capital ratios to prior changes in capital, interest and non–interest expense ratios, nonperforming loans, total assets, and other controls.
- Market power test:
  - Uses Panzar and Rosse H–statistic approach as implemented by Claessens and Laeven.
  - H–statistic equals sum of elasticities of bank revenue (interest revenue/ assets) with respect to input prices (interest and noninterest expense ratios).
  - Interpretation: H = 1 for perfect competition; H < 0 for monopoly; 0 < H < 1 for monopolistic competition.
  - The time period included a sharp decline in the short–term interest rate, testing how revenues responded.

### Estimation sample, variables, and controls
- Sample period for PR regressions: 2007 Q3 to 2008 Q4 (annual growth rates to avoid seasonality; first observation growth 2006 Q3–2007 Q3).
- Sample size: depending on specification, largest 79 to 87 bank holding companies with available data.
- Key control variables: initial off–balance sheet items to total assets, liquid assets to total assets, TARP funds to total assets (nonzero only in 2008 Q4), log of total assets, size dummies (based on average assets over 2006–2008), period dummies.
- Endogeneity handling: one–year lagged change in capital–to–asset ratio used; IV estimation performed (instruments: one–year lagged capital ratios and lagged profits or ROE) with similar findings.

### Main empirical findings
- Capital as a constraint:
  - Coefficients on lagged change in capital–asset ratio and its interaction with initial capital are positive and negative respectively, and both are statistically significant for both loan and deposit growth.
  - Results robust to measure of capital used (Total capital, Tier 1, or equity–asset ratio).
  - Stronger relationships during the capital loss period (2006 Q4–2008 Q4) than in the extended sample including 2009 H1.
  - Conclusion: capital, rather than liquidity (as measured by initial liquid asset ratio), was the primary constraint on lending during the crisis period analyzed.
- Liquidity:
  - Initial liquid asset ratio did not significantly predict higher subsequent deposit or loan growth.
  - Caveat: liquid asset ratio may poorly capture funding stress caused by interbank market disruptions and extensive use of repo funding.
- TARP effectiveness:
  - Ratio of TARP funds to total assets was not significant in regressions; TARP funds distributed in 2008 Q4 did not have an observable impact on loan supply in the short sample window analyzed (including one–quarter lags).
  - Consistent evidence from other work: of every new dollar of TARP capital injections (CPP flows), only between 10 and 18 cents went toward new lending, whereas about 67 cents went to boosting banks’ Tier 1 capital ratios.
- Off–balance sheet items:
  - Mixed results; the lagged ratio of off–balance sheet items to total assets is positive for loan growth in the capital loss period but negative in other tests; often statistically insignificant.
- Choice of bank capital (CC model tests):
  - Banks increased capital after past capital losses; the effect is larger for banks that were more capital constrained initially.
  - Quarterly bank stock return is positively related to Tier 1 and total capital ratios (sample where available).
  - Interest expense significant; non–interest expense and nonperforming loans show expected sign patterns with varying significance.
  - Time dummies indicate a generalized uptick in regulatory bank capital in 2008 Q4 that persisted into H1 2009 (partly due to TARP and balance sheet risk–weight reductions).
  - Model explains at least 63 percent of variation in total bank capital (specifications excluding quarterly stock return).
- Monopoly power / H–statistic:
  - Interest and noninterest expense ratios positively affect lending rates; nonperforming loans not significant.
  - Interaction terms indicate asymmetric responses when capital constraint binds versus when it does not: response of lending rates to noninterest expenses virtually disappears when capital constraint binds.
  - Tests reject H = 0 and H = 1 (F–statistics ≥ 6, p–values = 0.000), implying the sum of coefficients lies between zero and one—consistent with monopolistic competition among the largest bank holding companies.

### Conclusions and policy implications
- Summary conclusions:
  - Capital constraints were the principal factor limiting deposit and loan growth for large bank holding companies during the 2007–2009 crisis; liquidity (as measured) was not.
  - Banks optimally manage capital consistent with the CC model: demand for capital increases when initial capital is low and when costs are lower, and the response is convex.
  - Banking market structure for large bank holding companies is best characterized by monopolistic competition, enabling them to increase net interest margins when input costs fell.
  - Bank lending is unlikely to expand appreciably until banks complete rebuilding their capital positions.
- Policy recommendations concerning privatization of excess Federal Reserve assets:
  - The Federal Reserve acquired $1,142 billion of additional assets held by the shadow banking system during the crisis.
  - Privatization of these assets should avoid replaying the crisis; assets should be sold to private intermediaries subject to prudential regulation.
  - To privatize without destabilizing banks, $114.2 billion in total capital—10 percent of the required balance sheet expansion—must be added to the banking system so that institutions are "well qualified" under current commercial bank regulations.
  - Banks must receive clear signals that they can acquire these assets at a competitive risk–adjusted rate of return to incentivize optimal capital raising without excessive risk-taking.
  - Higher loan rates required to fund these assets would help mitigate excessive borrower risk.

*Italicized source attribution: IMF working paper content unit provided.*

### REFERENCES

### REFERENCES

### Bibliographic entries
- Bayazitova, Dinara and Anil Shivdasani, 2009, “Assessing TARP,” (unpublished: Chapel Hill, North Carolina: University of North Carolina).  
- Brunnermeier, M., 2009, “Deciphering the Liquidity and Credit Crunch 2007–2008,” Journal of Economic Perspectives, Vol. 23, No. 1, pp. 77–100.  
- Cecchetti, Stephen G., 2009, “Crisis and Responses: The Federal Reserve in the Early Stages of the Financial Crisis”, Journal of Economic Perspectives, Vol. 23, No. 1, pp. 51–75.  
- Cecchetti, Stephen G., 2008, “Monetary Policy and the Financial Crisis of 2007–08”, CEPR Policy Insight No. 21, April 2008.  
- Cetorelli, N., and M. Gambera, 2001, “Banking Market Structure, Financial Dependence and Growth: International Evidence from Industry Data”, The Journal of Finance, Vol. 56, No. 2, pp. 617–648.  
- Chailloux, Alexandre, 2009, “The Rise and Fall of the Repo Empire” forthcoming IMF working paper.  
- Chami, R., and T. Cosimano, 2001, “Monetary Policy with a Touch of Basel,” IMF Working Paper 01/151 (Washington: International Monetary Fund).  
- Chami, R., and T. Cosimano, 2010, “Monetary Policy with a Touch of Basel,” forthcoming Journal of Economics and Business.  
- Chari, V.V., Christiano, L., and P. J. Kehoe, 2009, “Facts and Myths about the Financial Crisis of 2008.” Federal Reserve Bank of Minneapolis Working Paper.  
- Claessens, S. and L. Laeven, 2004, “What Drives Bank Competition? Some International Evidence,” Journal of Money, Credit and Banking, Vol. 36, No. 3, pp. 563–583.  
- Cochrane, J. H., 2009, “Understanding Fiscal and Monetary Policy in 2008–2009”, University of Chicago Working Paper.  
- Diamond, D. W. and R. G. Rajan, 2009, “The Credit Crisis: Conjectures about Causes and Remedies,” University of Chicago Working Paper.  
- Gao, P., and H. Yun, 2009, “Commercial Paper, Lines of Credit, and the Real Effects of the Financial Crisis of 2008: Firm–Level Evidence from the Manufacturing Industry”, Documents the decline of CP and Increase in Line of Credit Working Paper June 2009.  
- Gorton, G., 2008, “The Subprime Panic,” NBER Working Paper 2008.  
- Gorton, G., 2009, “Slapped in the Face by the Invisible Hand: Banking and the Panic of 2007”, Paper prepared for the Federal Reserve Bank of Atlanta 2009 Financial Markets Conference.  
- Gorton, G. and A. Metrick, 2009, “Securitized Banking and the Run on Repo”, Yale ICF Working Paper, July 2009.  
- Hull, John C., 2006, Options, Futures, and Other Derivatives, (Upper Saddle River, New Jersey: Pearson Education International, 6th ed.).  
- Ivashina, V. and D. Scharfstein, 2009, “Bank Lending During the Financial Crisis of 2008,” EFA 2009 Bergen Meetings Paper.  
- Keister, Todd and James McAndrews, 2009, “Why are Banks Holding So Many Excess Reserves?,” Current Issues in Economics and Finance, Federal Reserve Bank of New York, Vol. 1, No.8. pp. 1–10.  
- Koch, T. and S. MacDonald, 2007, Bank Management (Marson, Ohio: South–Western Cengage Learning, 7th ed.).  
- Madigan, B., 2009, “Bagehot’s Dictum in Practice: Formulating and Implementing Policies to Combat the Financial Crisis”, Speech at Federal Reserve Bank of Kansas City Annual Economic Symposium, August.  
- Monetary Policy Report to the Congress, Board of Governors of the Federal Reserve System, February 2009.  
- Panzar, J. C. and J. N. Rosse, 1987, “Testing for ‘Monopoly Power’ Equilibrium,” Journal of Industrial Economics, Vol. 35, No. 4, pp. 443–456.  
- Peek, J., and E. Rosengren, 1995, “The Capital Crunch: Neither a Borrower nor a Lender Be,” Journal of Money, Credit and Banking, Vol. 27, No. 3, pp. 625–638.  
- Sarkar, Asani, 2009, “Liquidity Risk, Credit Risk, and the Federal Reserve’s Responses to the Crisis,” Federal Reserve Bank of New York, Staff Report no. 389.  
- Taliaferro, Ryan, 2009, “How do Banks Use Bailout Money? Optimal Capital Structure, New Equity, and the TARP,” (unpublished: Harvard Business School).

### Tables and regression material referenced
- Table 1. Selected Banking Indicators by Groups of Banks (percentages)  
- Table 2. Selected Banking Indicators by Size of Bank Holding Company  
- Table 3. Selected Banking Indicators for Bank Holding Companies that Received TARP Funding in 2008Q4  
- Table 4. Determinants of the Growth Rate of Bank Loans and Deposits  
- Table 5. Regressions for the Choice of Bank Capital

### Table 6. Test of Monopoly Power (time period: 2006 Q1 - 2008 Q4)
- Dependent variable: Interest income ratio  
- Explanatory variables and OLS regression coefficients (robust standard errors in parentheses):
  - Interest expense ratio 0.51  * * * (0.07)  
  - Interest expense ratio · Capital Constraint Dummy 0.07 (0.119)  
  - Noninterest expense ratio 0.09 *** (0.029)  
  - Noninterest expense ratio · Capital Constraint Dummy -0.09 (0.082)  
  - Ratio of nonperforming loans to assets 0.00 (0.000)  
  - Nonperforming loan ratio · Capital Constraint Dummy 0.00 (0.001)  
  - Capital Constraint Dummy -0.07 (0.176)  
  - Logarithm of assets -0.01 (0.014)  
  - Dummy =1 for banks whose assets >$17 billion and <$460 billion -0.02 (0.062)  
  - Dummy =1 for banks whose assets <$17 billion 0.06 (0.080)  
  - Dummy for 2007Q3 0.23  * * * (0.029)  
  - Dummy for 2007Q4 0.18  * * * (0.029)  
  - Dummy for 2008Q1 0.12  * * * (0.025)  
  - Dummy for 2008Q2 0.08  * * * (0.023)  
  - Dummy for 2008Q3 0.07  * * * (0.026)  
  - Constant 0.84  * * * (0.138)  
  - R-squared 0.5746
- Tests on the coefficients: F-statistic (p-value)
  - H0: Sum of the coefficients on input prices = 0  
    - Capital constraint not binding 45.87 (0.000)  
    - Capital constraint binding 12.26 (0.000)
  - H0: Sum of the coefficients on input prices = 1  
    - Capital constraint not binding 20.89 (0.000)  
    - Capital constraint binding 6.56 (0.010)
- Notes on the regression:
  - This table shows the results of OLS regressions for the interest income ratio to total assets. Other explanatory variables include the interest and noninterest expense ratios, as well as their interactions with a capital constraint dummy which takes the value of 1 if the total capital ratio is below 10 percent, and zero otherwise. The logarithm of total assets, dummy variables distinguishing banks according to three size categories are included, and time dummy variables are included as regressors. Robust standard errors are shown in parentheses, and significance levels of 10 (*), 5 (**), and 1 (***) percent are indicated.

*Source: _wp10131 - REFERENCES*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp10131.pdf_