## _wp10278 — References (Selected Content Unit)

## Source details

**Canonical URL:** [_wp10278 — References (Selected Content Unit)](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp10278.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp10278.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp10278.pdf.json)

---

### Introduction: question, theory, and approach
- Research question: Do U.S. monetary policy shocks disproportionately affect stock returns of firms that are more dependent on external finance?
- Theoretical channels:
  - Balance sheet channel: interest-rate changes alter collateral values, firm net wealth, and the premium for external finance.
  - Lending channel: monetary policy affects banks’ credit supply, changing cost and quantity of borrowing.
- Cross-sectional prediction: stock price reaction to monetary policy shocks varies across firms by financial dependence.
- Identification strategy:
  - Exogenous U.S. monetary policy shocks measured as the one-day change in the price of the one-month federal funds futures contract on FOMC announcement days (Bernanke and Kuttner (2005) approach).
  - Industry financial dependence measured using U.S. data following Rajan and Zingales (1998).
  - U.S. firms dropped from the sample to strengthen exogeneity for foreign firms.
- Advantage: uses high-frequency stock prices for event-study analysis and controls for unobserved time-invariant effects via cross-sectional asymmetric identification.

### Methodology (empirical strategy and key equations)
- Main panel regression specification:
  - Stock Returnijkt = β Monetary Shockt + γ Controlit + εijkt  (notation: i = company, j = country, k = sector, t = time).
  - Firm fixed effects included; standard errors clustered at FOMC date level.
- Heterogeneity and interaction specifications:
  - βk = β1 + β2 Financial Dependencek  — β2 measures dependence heterogeneity.
  - βkt = β1 + β2 Financial Dependencek + β3 Financial Dependencek * Recessiont  — Recessiont uses NBER recession quarters.
  - βkjt = β1 + β2 Financial Dependencek + β4 Country Traitjt  — Country Trait examples: financial integration and financial development.
- Controls:
  - Fama and French (1992) controls: firm size (log assets), market-to-book ratio, and beta*market return.
  - Firm characteristics lagged one year except contemporaneous stock market return.
  - Standard errors clustered at the time level for interaction specifications.

### Data and variable construction
- Sample coverage:
  - 20,121 firms in 44 countries.
  - Period: 1990 to 2008.
  - Total FOMC meetings: 140.
  - Total firm-date observations: 925,306.
- Dependent variable:
  - Two-day stock return: log difference in closing price over t-1 and t+1 (t = FOMC meeting day).
  - Two-day returns above 50% or below -50% dropped (covers 0.1% of sample).
  - Robustness: winsorize top and bottom 1% — main results robust.
- Monetary policy shock:
  - Measured as change in current-month federal funds futures rate scaled by days-in-month factor (Kuttner (2001); Bernanke and Kuttner (2005)).
  - Extended coverage through 2008; include FOMC dates with no policy rate change as control dates.
- Financial dependence index:
  - Financial dependence = (Capital expenditures - Cash flow) / Capital expenditures (Rajan and Zingales (1998)).
  - Cash flow = cash flow from operations + decreases in inventories + decreases in receivables + increases in payables.
  - Recomputed for 1990-2006 for ~150 SIC 3-digit sectors; drop SIC 4, 5, 6, and 9 sectors.
  - Sector-level median computed for SIC 3-digit sectors with at least 5 firm observations.
- Recession dating: NBER recession dates used to construct Recession dummy for FOMC meeting dates within recession quarters.

### Key empirical findings (select results and quantitative illustrations)
- Average effect sign:
  - U.S. monetary policy loosening (tightening) associated with increases (decreases) in foreign stock prices.
- Financial dependence heterogeneity:
  - Firms with higher intrinsic external finance dependence exhibit larger stock price responses to U.S. monetary shocks.
  - Illustrative magnitude: an unexpected policy rate decrease of 5 basis points (equal to its interquartile range) is associated with a stock price response that is 6 basis points greater for firms at the 75th percentile of financial dependence (Construction machinery) relative to firms at the 25th percentile (Beverages).
  - Context: average stock market return around FOMC dates is 19 basis points.
- Business-cycle asymmetry:
  - Effect stronger during U.S. recession periods; financing channel more important during recessions (β3 expected to be negative).
- Cross-country heterogeneity:
  - Impact more pronounced in countries with greater financial integration (foreign assets and liabilities over GDP, Lane and Milesi-Ferretti (2007)).
- Descriptive ranges and sample statistics (selected exact values):
  - Actual federal funds rate changes announced: Min -100 bp; Max 75 bp.
  - Unexpected rate shocks (Shock): Min -43 bp; Max 24 bp.
  - Financial dependence index: Min -2.4; Max 1.4.
  - Stock return (in %): Obs 925306; Mean 0.19; Median 0; St Dev. 5.38; P25 -1.79; P75 2.00; Min -50; Max 50.
  - Shock (in bp): Obs 140; Mean -3.51; Median 0; St Dev. 10.63; P25 -5; P75 0; Min -43; Max 24.
  - Financial Openness: Obs 564; Mean 2.38; Median 1.41; St Dev. 2.53; P25 0.96; P75 2.96; Min 0.3; Max 23.9.
  - Trade Openness: Obs 564; Mean 0.02; Median 0.01; St Dev. 0.07; P25 -0.02; P75 0.05; Min -0.2; Max 0.3.
- Estimated magnitudes (selected reported coefficients and significance):
  - Baseline Shock coefficients: -0.043**, -0.041**, -0.038**, -0.038**, -0.013** (various specifications).
  - Shock*Financial dependence coefficients (Table 2 and robustness): -0.017***, -0.018***, -0.012***, -0.011***.
  - Table 3 (clustering robustness) Shock*Financial dependence = -0.011*** with reported standard errors [0.0035], [0.0023], [0.0029], [0.0009] across clustering choices.
  - Table 6 (business cycle) Recession*Shock*Financial dependence reported -0.014*** [0.0046].

### III. EMPIRICAL RESULTS — Basic specification and robustness
- Baseline sample and fixed effects: firm fixed effects; standard errors clustered at FOMC meeting dates.
- Decomposition results (Column highlights):
  - Column 1: actual federal funds rate changes → negative coefficient but not significantly different from zero in many specs.
  - Column 2: expected vs. unexpected components — unexpected component significantly negative; estimated coefficient 0.04 (interpretation: a 25-base point increase reduces global stock prices by about 1%).
  - One-standard deviation increase in the monetary shock → reduces stock prices by 0.4%; explains 8% of the standard deviation of stock returns.
  - Expected rate change component: positive coefficient, less statistically significant; economic impact about 10% of unexpected shock impact.
- Interaction and controls:
  - Including Shock * Financial dependence yields statistically significantly larger impact for sectors more dependent on external finance.
  - Adding firm size and market-to-book increases coefficient on Shock*Financial dependence somewhat; adding beta reduces Shock*Financial dependence from 0.018 to 0.012 but remains significant at the 1% level.
  - Including FOMC date dummies: Shock variable drops; Shock*Financial dependence remains significant with coefficient -0.011.

### III.B Robustness checks (summary)
- Standard-error clustering: Shock*Financial dependence remains significant at 1% across clustering at country, 3-digit SIC industry, country-year, and two-way date and industry clusters.
- FOMC meetings without rate change: excluding dates with no rate change (Table 4) → results similar to baseline.
- Manufacturing-only sample: restricting to manufacturing (Table 5) → Shock*Financial dependence remains statistically significant.
- Reverse causality and simultaneity:
  - Evidence and prior literature (Bernanke and Kuttner (2005)) suggest limited systematic contemporaneous monetary response to stock prices; potential reverse causality would bias estimates toward zero, not spuriously produce significant negative effects.

### III.D Growth opportunity and business cycle
- Alternative channels tested:
  - Industry lagged sales growth and its interaction with Financial dependence included — interaction enters with negative coefficient; main Shock*Financial dependence result unchanged.
- Business cycle asymmetry:
  - Recession dummy based on NBER dates: July 1990 to March 1991; March 2001 to November 2001; and December 2007 till 2008.
  - Interaction results: disproportionate effect of U.S. monetary shocks on financially dependent firms significantly larger during recessions.

### III.E Cross-country variations and spillovers
- Financial openness:
  - Shock * Financial openness interaction significantly negative at the 5% level in some specs and at 1% in combined specs (examples: Financial Openness coefficient -0.057*; Shock*Financial Openness reported -0.0018** and -0.0021*** with standard errors [0.00073], [0.00065]).
  - Main Shock * Financial dependence coefficient little changed when controlling for openness.
- Domestic financial development and trade openness:
  - Interaction with shock not significant for domestic credit to private sector over GDP or trade openness measures in multiple specifications.
- Interest-rate synchronization:
  - Monthly correlation of money market rates (1990–2008) used; high-synchronization defined as correlation > 0.5.
  - High Synchronization*Shock reported -0.018*** [0.0070]; High Synchronization*Shock*Financial dependence reported -0.011*** [0.0041].
- Summary: spillovers of U.S. monetary shocks more important for more financially integrated countries and for countries with higher money-market-rate synchronization with the U.S.

### Robustness and inference (selected statistics)
- Firm fixed effects in all regressions; date fixed effects included in many specifications.
- R-squared ranges reported:
  - Table 2: 0.025 to 0.092.
  - Table 4 (restricted): 0.047 to 0.115.
  - Table 5 (manufacturing): 0.027 to 0.104.
  - Table 6: 0.092.
  - Table 7: 0.092 to 0.094.
- Observations vary by specification (examples):
  - Full-sample observations: up to 925,306.
  - Subsample (excluding no-change FOMC dates): Observations 473,713 in Table 4 examples.
  - Table 7 specification observations: 871,415; 825,335; 862,522; with R-squared around 0.092–0.094.

### IV. Conclusions (concise)
- Dataset: 20,121 firms across 44 countries over 1990-2008.
- Main empirical findings:
  - Stock prices increase following unexpected monetary loosening and decrease following unexpected monetary tightening.
  - Effects are larger for sectors dependent on external financing, especially during recessions, and in countries more integrated with global finance.
- Identification advantage: asymmetric cross-sectional strategy controls for unobserved time-invariant effects and helps alleviate endogeneity and simultaneity concerns.
- Economic significance (reiterated example):
  - An unexpected policy rate decrease of 5 basis points (interquartile range) → associated with a stock price response 6 basis points greater for firms at the 75th percentile of financial dependence relative to the 25th percentile; average stock market return around FOMC dates is 19 basis points.
- Policy implications:
  - U.S. monetary policy has cross-border spillovers that operate through financial-frictions channels, affecting global asset allocation and the dollar value of equity in financially dependent sectors and in more financially integrated countries.

*Source: _wp10278 — 3. Correlation between U.S. and Local Money Market Rates; 0.3 percent for the Republic of Korea in year 1991 to a high of 23.9 percent for Hong Kong; Appendix Tables and References (from _wp10278.pdf)*

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

### _wp10278 - References .............................................................................................................

### References
- References ..................................................................................................................................... 16 

### Tables
- 1. Summary Statistics ................................................................................................................. 19
- 2. The Effect of Monetary Policy Shocks on Stock Returns ..................................................... 20
- 3. Clustering of Standard Errors ................................................................................................ 21
- 4. Excluding FOMC Dates Without Change in Federal Funds Rate ......................................... 22
- 5. The Impact of Monetary Policy Shocks on Manufacturing Sectors ...................................... 23
- 6. Growth Prospects and Business Cycles ................................................................................. 24
- 7. Spillover Channels ................................................................................................................. 25

### Appendix
- 1.   List of Countries .................................................................................................................... 26
- 2. List of FOMC Meeting, Actual Change in Federal Funds Rate,  
        and Monetary Shock, in bp .................................................................................................. 27

*Source: _wp10278 - References .........................................................................................................*

### 3. Correlation between U.S. and Local Money Market Rates ................................................... 28

### 3. Correlation between U.S. and Local Money Market Rates ................................................... 28

### Introduction: question, theory, and approach
- Research question: Do U.S. monetary policy shocks disproportionately affect stock returns of firms that are more dependent on external finance?
- Theoretical channels emphasized:
  - Balance sheet channel: changes in interest rates alter value of collateral, firm net wealth, and premium paid for external finance.
  - Lending channel: monetary policy affects banks’ credit supply, changing cost and quantity of borrowing for firms.
- Cross-sectional prediction: stock price reaction to monetary policy shocks should vary across firms by financial dependence.
- Identification strategy:
  - Use exogenous U.S. monetary policy shocks measured as the one-day change in the price of the one-month federal funds futures contract on FOMC announcement days (following Bernanke and Kuttner (2005)).
  - Measure industry financial dependence using U.S. data following Rajan and Zingales (1998).
  - Drop U.S. firms from the sample to strengthen exogeneity of U.S. shocks for foreign firms.
- Advantages of the approach:
  - Controls for unobserved time-invariant effects via cross-sectional asymmetric identification.
  - Uses high-frequency stock prices for event-study analysis to isolate short-term responses.

### Methodology (empirical strategy and key equations)
- Main panel regression specification (notation from the source):
  - Stock Returnijkt = β Monetary Shockt + γ Controlit + εijkt  (equation (1))
  - i = company, j = country, k = sector, t = time. Firm fixed effects included; standard errors clustered at FOMC date level.
- Interaction to capture dependence heterogeneity:
  - βk = β1 + β2 Financial Dependencek  (equation (2))
  - β2 measures how the effect of the monetary shock depends on an industry’s external financing needs.
- Business-cycle asymmetry test:
  - βkt = β1 + β2 Financial Dependencek + β3 Financial Dependencek * Recessiont  (equation (3))
  - Recessiont is a dummy for FOMC meeting dates occurring during NBER recession quarters; β3 captures additional effect during recessions.
- Cross-country heterogeneity:
  - βkjt = β1 + β2 Financial Dependencek + β4 Country Traitjt  (equation (4))
  - Country Trait examples: financial integration and financial development; β4 captures dependence on country traits.
- Controls and estimation details:
  - Include three Fama and French (1992) factors: firm size (log assets), market-to-book ratio, and beta*market return.
  - Firm characteristics lagged by one year except contemporaneous stock market return.
  - Cluster standard errors at the time level for interaction specifications.

### Data and variable construction
- Sample:
  - 20,121 firms in 44 countries.
  - Period: 1990 to 2008.
  - Total FOMC meetings: 140.
  - Total firm-date observations: 925,306.
- Dependent variable: two-day stock return computed as the log difference in closing price over t-1 and t+1 (t = FOMC meeting day).
  - Two-day window chosen to account for time zone differences.
  - Extreme two-day returns above 50% or below -50% dropped (covers 0.1% of sample).
  - Robustness: winsorize top and bottom 1% — main results robust.
- Monetary policy shock:
  - Measured following Kuttner (2001) and Bernanke and Kuttner (2005): change in current-month federal funds futures rate scaled by days-in-month factor so Shock = (d - 1)/D * (fd - fd-1) interpretation per source notation.
  - Extended coverage of shocks through 2008 and include FOMC dates with no policy rate change as control dates.
  - Rationale: daily futures-change measure abstracts from anticipated actions.
- Financial dependence index:
  - Based on Rajan and Zingales (1998): Financial dependence = (Capital expenditures - Cash flow) / Capital expenditures.
  - Cash flow = cash flow from operations + decreases in inventories + decreases in receivables + increases in payables.
  - Recomputed for 1990-2006 to expand coverage to ~150 SIC 3-digit sectors; drop SIC 4 (utilities), SIC 5 (wholesale and retail), SIC 6 (financial), and SIC 9 (public administration).
  - Sector-level median of firm ratios computed for SIC 3-digit sectors with at least 5 firm observations.
- Recession dating:
  - Use NBER recession dates (quarterly frequency) to construct Recession dummy for FOMC meeting dates within recession quarters.
- Control variables:
  - Firm size = log total assets (Worldscope/Datastream).
  - Market-to-book = market value of equity / book value of equity.
  - Firm beta = correlation between firm weekly return and weekly country-level stock market return (domestic beta used).

### Key empirical findings (select results and quantitative illustrations)
- Sign of average effect:
  - Strong evidence of a negative response of stock prices to U.S. monetary shocks: U.S. monetary policy loosening (tightening) associated with increases (decreases) in foreign stock prices.
- Financial dependence heterogeneity:
  - Impact is particularly pronounced for firms with relatively high intrinsic dependence on external finance.
  - Illustrative magnitude: an unexpected policy rate decrease of 5 basis points (equal to its interquartile range) is associated with a stock price response that is 6 basis points greater for firms at the 75th percentile of financial dependence (Construction machinery industry) relative to firms at the 25th percentile (Beverages industry).
  - Context: average stock market return around FOMC dates is 19 basis points.
- Business-cycle asymmetry:
  - Effect is stronger during U.S. recession periods, consistent with financial constraints being more binding in downturns (β3 expected to be negative).
- Cross-country heterogeneity:
  - Impact of U.S. monetary shocks on stock prices is more pronounced in countries that are more financially integrated with the rest of the world.
  - Financial integration measured as a country’s foreign assets and liabilities over GDP (following Lane and Milesi-Ferretti (2007)).
- Descriptive ranges and sample statistics:
  - Actual federal funds rate changes announced range from a rate cut of 100 bps to a rate increase of 75 bps.
  - Unexpected rate shocks range from -43 bps to +24 bps.
  - Financial dependence index ranges from -2.4 (Manifold Business Forms) to 1.4 (Photographic Equipment and Supplies).
  - Sample winsorization/drop rules: two-day returns outside ±50% dropped (0.1% of sample); winsorization at 1% used as robustness.

### Interpretation and implications
- Evidence supports the role of financial frictions in the transmission of monetary policy to asset prices: firms more dependent on external finance experience larger stock price responses to U.S. monetary shocks.
- U.S. monetary policy influences global asset allocation and the cross-border transmission of shocks, with stronger effects where countries are more financially integrated.
- Policy-relevant points:
  - Cross-border spillovers matter for asset prices; domestic monetary policy in the U.S. can have real effects abroad via financial-frictions channels.
  - The business-cycle state alters the strength of transmission through financial dependence, implying countercyclical sensitivity of financially dependent firms during recessions.

*Source: _wp10278 - 3. Correlation between U.S. and Local Money Market Rates; https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp10278.pdf*

### 0.3 percent for the Republic of Korea in year 1991 to a high of 23.9 percent for Hong Kong in

### _wp10278 - 0.3 percent for the Republic of Korea in year 1991 to a high of 23.9 percent for Hong Kong in

### III. EMPIRICAL RESULTS — Basic Specification
- Baseline sample: firm fixed effects throughout; cluster standard errors at the level of FOMC meeting dates.
- Column 1: actual federal funds rate changes → negative coefficient on actual federal funds rate variable but not significantly different from zero.
- Column 2: federal funds rate decomposed into expected and unexpected components (Bernanke and Kuttner (2005) method).
  - Unexpected component: significantly negative coefficient.
  - Estimated coefficient: 0.04.
  - Interpretation: a 25-base point increase of U.S. rate would reduce global stock prices by about 1%.
  - One-standard deviation increase in the monetary shock → reduces stock prices by 0.4%; explains 8% of the standard deviation of stock returns.
  - Expected rate change component: positive coefficient, less statistically significant; economic impact about 10% of the unexpected rate shock impact.
- Column 3: excluding expected rate shocks → unexpected rate shock coefficient similar to Column 2 (unexpected and expected shocks orthogonal).
- Column 4: asymmetric impact by sector-level dependence on external finance (Rajan and Zingales (1998) financial dependence).
  - Interaction Shock * Financial dependence included (financial dependence itself absorbed by firm dummies).
  - Finding: impact of interest rate shocks statistically significantly higher for sectors more dependent on external finance.
  - Quantitative example: an unexpected policy rate decrease of 5 basis points (equal to its interquartile range) is associated with a stock price response that is 6 basis points greater for firms at the 75th percentile of financial dependence (Construction machinery) relative to the 25th percentile (Beverages).
  - Average stock market return around FOMC dates: 19 basis points.
- Column 5: include firm size and market to book ratio (Fama and French (1992) controls) → increases coefficient on Shock*Financial dependence somewhat; statistical significance almost the same.
- Column 6: control for beta factor (interaction between firm’s beta and local stock market return); beta lagged one-year, computed as correlation of weekly individual stock returns and local market return within a year.
  - Adding beta reduces Shock*Financial dependence coefficient from 0.018 to 0.012; remains significant at the 1% level.
- Last column: include dummies for FOMC dates → Shock variable drops; Shock*Financial dependence remains significant with coefficient -0.011, comparable to results without FOMC date effects.

### III.B Robustness Checks of Main Results
- Clustering of standard errors:
  - Table 3: model same as last baseline column but vary clustering.
  - Column 1: cluster at country level.
  - Column 2: cluster at 3-digit SIC industry level.
  - Column 3: cluster at interaction of country and year.
  - Column 4: two-way clustering by date and industry.
  - Across specifications: Shock*Financial dependence consistently significant at the level of 1 percent.
- FOMC meetings without rate change:
  - Original sample includes FOMC meetings with no change in Federal Funds rate.
  - Robustness check drops cases with no rate change (Table 4) → results similar to baseline.
- Manufacturing sector only:
  - Rajan and Zingales (1998) financial dependence was computed for manufacturing only.
  - Exclude utility, trade, financial, and government sectors already.
  - Re-estimate on manufacturing subset (Table 5) → Shock*Financial dependence remains statistically significant at conventional levels.

### III.C Reverse Causality and Simultaneity Bias
- Concern: monetary policy responding contemporaneously to stock market would bias results.
  - Bernanke and Kuttner (2005) do not find evidence of systematic reaction.
  - If FOMC cut rates in response to large drops in equity prices, reverse causality would reduce estimated response and bias toward finding no effect.
- Joint determination (monetary policy and stock prices responding to same news) could create simultaneity bias.
  - Example: bad news → stock declines and loosening of policy → tendency for rate cuts to be associated with stock market declines → downward bias in estimated response.
  - Such biases would attenuate estimated effects, mitigating concern that significant results are driven by reverse causality/simultaneity.

### III.D Growth Opportunity and Business Cycle
- Test alternative channels (demand/growth opportunities) and asymmetries across business cycle.
- Column 1 of Table 6:
  - Control for industry’s lagged sales growth (U.S. data proxy for growth opportunities, following Fisman and Love (2007)) and its interaction with Financial dependence.
  - Interaction enters with negative coefficient → suggests monetary shocks may also affect stock prices through demand channel.
  - Main Shock*Financial dependence result unchanged.
- Business cycle asymmetry:
  - Financial constraints expected to bind more during recessions (Braun and Larrain (2005)).
  - Recession dummy based on NBER recession dates: July 1990 to March 1991; March 2001 to November 2001; and December 2007 till 2008 (end of sample period).
  - Columns 2 and 3: include interaction between Shock * Financial Dependence and Recession dummy.
  - Finding: disproportionate effect of U.S. monetary shocks on financially dependent firms is significantly larger during economic recessions — financing channel more important during recessions.

### III.E Cross-Country Variations and Spillovers
- Table 7: examine cross-country heterogeneity; include country fixed effects and FOMC meeting fixed effects.
- Financial openness:
  - Measure: country’s foreign assets and liabilities over GDP (Lane and Milesi-Ferretti (2007)).
  - Interaction Shock * Financial openness: significantly negative at the 5% level → spillovers of U.S. shocks more important for more financially integrated countries.
  - Main Shock * Financial dependence coefficient little changed.
- Domestic financial development:
  - Measure: domestic credit to private sector over GDP.
  - Interaction with shock not significant.
- Trade openness:
  - Measure: imports plus exports over GDP.
  - Interaction with shock not significant.
- Combined channels:
  - Column 4: include financial openness and trade channels together → Shock * Financial Openness significant at 1%; Shock * Financial dependence remains significant at 1%.
- Economic development:
  - Column 5: include log per capita GDP and its interaction with shock → neither significant; Shock * Financial dependence retains magnitude and significance.
- Interest rate synchronization:
  - Measure: correlation of monthly money market rates between U.S. and individual countries over 1990–2008 (for Euro countries use Euro money market rate after joining Euro).
  - Define high-synchronization if correlation coefficient > 0.5 (Appendix Table 3 lists coefficients).
  - Column 6: asymmetric impact more pronounced in countries with high synchronization with U.S. monetary policy.
- Summary: strong and robust asymmetric relationship between stock responses of financially dependent firms and U.S. monetary policy shocks.

### IV. CONCLUSIONS
- Dataset: 20,121 firms across 44 countries over 1990-2008.
- Main empirical findings:
  - Stock prices tend to increase following unexpected monetary loosening and decrease following unexpected monetary tightening.
  - Impact more pronounced for sectors dependent on external financing, especially during economic recessions, and for countries more integrated with the global financial market.
- Identification advantage:
  - Asymmetric cross-sectional strategy (differential effects across firms by external financial dependence) controls for unobserved time-invariant effects that could affect both monetary policy and firm stock prices, alleviating endogeneity and simultaneity concerns.
  - Extensions consider business cycle asymmetries to further distinguish financial channel from alternative explanations.
- Economic significance:
  - Example repeated: an unexpected policy rate decrease of 5 basis points (equal to its interquartile range) → associated with a stock price response 6 basis points greater for firms at the 75th percentile of financial dependence (Construction machinery) relative to 25th percentile (Beverages).
  - Compare to average stock market return around FOMC dates: 19 basis points.
- Implications:
  - Prices in stock markets respond strongly to monetary shocks.
  - Financial frictions play an important role in transmission of monetary policy.
  - U.S. monetary policy influences global capital allocation.

*Source: _wp10278 - 0.3 percent for the Republic of Korea in year 1991 to a high of 23.9 percent for Hong Kong in*

### References

### _wp10278 - References

### Literature cited
- Extensive bibliography on monetary policy transmission, financial dependence, and asset returns including works by:
  - Ashcraft and Campello (2007); Baker, Stein, and Wurgler (2003); Barakchian and Crowe (2010); Bekaert, Hoerova, and Lo Duca (2010).
  - Bernanke and collaborators: Bernanke and Blinder (1992); Bernanke and Gertler (1995, 2001); Bernanke and Kuttner (2005); Bernanke and Mihov (1998); Bernanke (2010).
  - Fama (1981); Fama and Schwert (1977); Fama and French (1992).
  - Rajan and Zingales (1998); King and Levine (1993); Kroszner, Laeven, and Klingebiel (2007).
  - Mishkin (1996, 2009); Piazzesi and Swanson (2008); Rudebusch (1998); Taylor (2007).
- Methodological and empirical sources on measurement of shocks and futures-based shock identification: Bernanke and Kuttner (2005); Romer and Romer (2004); Kuttner (2001); Piazzesi and Swanson (2008).

### Data, variable definitions, and sample (from Table 1)
- Sample coverage:
  - 20,121 firms across 44 countries over the period 1990-2008.
  - Total of 925,306 firm-year observations.
- Key variable definitions:
  - Stock return: log difference in the closing price over the period t-1 and t+1, where t is the FOMC meeting date.
  - Actual change: actual change in federal funds rate announced at FOMC meetings.
  - Shock: change in the federal funds rate implied by the current-month futures contract from Bernanke and Kuttner (2005).
  - Expected change: difference between actual change and shock in federal funds rate.
  - Financial dependence: industry intrinsic dependence on external finance for investment based on Rajan and Zingales (1998).
  - Firm size: log of book assets in U.S. dollar.
  - Market beta: correlation between weekly firm stock price and the weekly country-level local market index within a year.
  - Financial openness: country’s foreign assets and liabilities over GDP.
  - Trade openness: country’s imports plus exports over GDP.
- Selected summary statistics (exact values):
  - Stock return (in %): Obs 925306; Mean 0.19; Median 0; St Dev. 5.38; P25 -1.79; P75 2.00; Min -50; Max 50.
  - Actual change (in bp): Obs 140; Mean -4.82; Median 0; St Dev. 26.29; P25 -25; P75 0; Min -100; Max 75.
  - Expected change (in bp): Obs 140; Mean -1.31; Median 0; St Dev. 22.19; P25 -8.5; P75 4.5; Min -92; Max 61.
  - Shock (in bp): Obs 140; Mean -3.51; Median 0; St Dev. 10.63; P25 -5; P75 0; Min -43; Max 24.
  - Financial dependence: Obs 150; Mean -0.06; Median 0.04; St Dev. 0.7; P25 -0.37; P75 0.33; Min -2.4; Max 1.4.
  - Shock*Financial dependence: Obs 19,375; Mean 0.09; Median 0; St Dev. 6.86; P25 -0.85; P75 0.74; Min -48.6; Max 61.7.
  - Firm size: Obs 924,690; Mean 12.00; Median 11.98; St Dev. 1.93; P25 10.81; P75 13.19; Min -1.16; Max 19.8.
  - Market/Book: Obs 899,602; Mean 2.40; Median 1.49; St Dev. 3.00; P25 0.85; P75 2.65; Min 0.2; Max 20.3.
  - Beta*Market Return: Obs 897,219; Mean 0.41; Median 0.11; St Dev. 1.88; P25 -0.27; P75 0.89; Min -26.1; Max 41.9.
  - Financial Openness: Obs 564; Mean 2.38; Median 1.41; St Dev. 2.53; P25 0.96; P75 2.96; Min 0.3; Max 23.9.
  - Trade Openness: Obs 564; Mean 0.02; Median 0.01; St Dev. 0.07; P25 -0.02; P75 0.05; Min -0.2; Max 0.3.

### Main empirical findings (from Tables 2–6)
- Baseline effect of monetary policy shock on stock returns:
  - Shock coefficient values reported (various specifications): -0.043**, -0.041**, -0.038**, -0.038**, -0.013** (robust standard errors reported).
  - Shock is negative and statistically significant in multiple specifications (significance levels indicated: ** p<0.05, *** p<0.01).
- Interaction with financial dependence:
  - Shock*Financial dependence coefficients: -0.017***, -0.018***, -0.012***, -0.011*** in Table 2 (standard errors reported).
  - Table 3 reports Shock*Financial dependence = -0.011*** across multiple clustering choices (Country level; Sector level; Country-year level; Country and sector level) with standard errors [0.0035], [0.0023], [0.0029], [0.0009] respectively.
  - Table 4 (excluding FOMC dates without change in FFR) reports Shock*Financial dependence = -0.017***, -0.012***, -0.012*** (standard errors provided).
  - Table 5 (manufacturing sectors) reports Shock*Financial dependence = -0.012***, -0.012***, -0.0061**, -0.0060** (standard errors provided).
  - Table 6 reports Shock*Financial dependence = -0.011***, -0.012***, -0.0054* across growth and recession specifications (standard errors provided).
- Other firm-level controls:
  - Market/Book ratio positive and significant: coefficients reported 0.045***, 0.032***, 0.029*** (Table 2) and 0.029*** repeatedly across robustness tables.
  - Beta*Market Return positive and significant: coefficients around 0.64***, 0.56***, 0.60***, 0.54***, 0.66***, 0.60*** in various specifications (standard errors reported).
  - Firm size coefficients variable and generally not significant across many specifications (examples: 0.067; -0.045; -0.032; -0.032; -0.033; -0.03).
- Components decomposition:
  - Actual change often small/insignificant (example reported coefficient -0.0029 with [0.0040]).
  - Expected change positive and sometimes significant (example 0.0052* [0.0028]; 0.0057** [0.0026]).
- Subsample and robustness results:
  - Excluding FOMC dates without change in FFR reduces sample to Observations 473,713 (Table 4) with Shock coefficients remaining negative and significant (examples -0.038**, -0.037**, -0.034**).
  - Manufacturing-only sample (Observations up to 606,315) shows negative Shock and negative Shock*Financial dependence with smaller magnitudes for interaction in some specs.
  - Clustering standard errors at different levels yields consistent Shock*Financial dependence estimates (Table 3).
- Growth prospects and business cycle interactions (Table 6):
  - Growth Opportunity coefficient reported 0.038 [0.11] (not significant given large standard error).
  - Shock*Growth Opportunity reported -0.031 [0.021].
  - Recession*Financial dependence coefficients reported -0.12 and -0.18** (standard errors as given).
  - Recession*Shock*Financial dependence reported -0.014*** [0.0046].
- Spillover channels (Table 7):
  - Shock*Financial dependence = -0.011*** reported across specifications with controls for Financial Openness, Financial Development, Trade Openness, Fin-trade Openness, GDP per capita, and Policy Rate Correlations (standard errors reported).
  - Financial Openness coefficient reported -0.057* and -0.051 (standard errors [0.032]).
  - Interaction Shock*Financial Openness reported -0.0018** and -0.0021*** (standard errors [0.00073], [0.00065]).
  - High Synchronization*Shock reported -0.018*** [0.0070] and High Synchronization*Shock*Financial dependence reported -0.011*** [0.0041].
  - Observations by specification: e.g., 871,415; 825,335; 871,415; 871,415; 871,415; 862,522. R-squared values around 0.092–0.094.

### Robustness and inference
- Standard errors clustered at the level of intervention dates in main tables; Table 3 demonstrates robustness to clustering at Country level, Sector level, Country-year level, and Country and sector level.
- Firm fixed effects included in all reported regressions; date fixed effects included in many specifications (reported per table).
- R-squared values vary by specification:
  - Table 2: R-squared range reported 0.025 to 0.092 across columns.
  - Table 4 (restricted sample): R-squared reported 0.047 to 0.115.
  - Table 5 (manufacturing): R-squared reported 0.027 to 0.104.
  - Table 6: R-squared reported 0.092 across columns.
  - Table 7: R-squared reported 0.092 to 0.094.

*Content unit: _wp10278 - References*

### Appendix Table 1. List of Countries

### _wp10278 - Appendix Tables (Tables 1–3)

### Appendix Table 1. List of Countries — Number of firms (1990–2008)
- Argentina 51
- Australia 1,701
- Austria 108
- Belgium 116
- Brazil 236
- Canada 1,662
- Chile 79
- China 1,387
- Colombia 25
- Czech Republic 43
- Denmark 135
- Egypt 39
- Finland 117
- France 844
- Germany 732
- Greece 223
- Hong Kong 611
- Hungary 23
- India 890
- Indonesia 227
- Ireland 71
- Israel 142
- Italy 249
- Japan 3,061
- Korea (South) 973
- Malaysia 781
- Mexico 83
- Netherlands 197
- New Zealand 94
- Norway 230
- Pakistan 88
- Peru 59
- Philippines 110
- Poland 205
- Portugal 78
- Russian Federation 60
- Singapore 474
- South Africa 415
- Spain 117
- Sweden 416
- Switzerland 186
- Thailand 391
- Turkey 176
- United Kingdom 2,216
- Total 20,121

### Appendix Table 2. List of FOMC Meetings, Actual Change in Federal Funds Rate, and Monetary Shock, in bp
- 13-Jul-90 — Actual -25 — Shock -14
- 29-Oct-90 — Actual -25 — Shock -31
- 14-Nov-90 — Actual -25 — Shock 4
- 7-Dec-90 — Actual -25 — Shock -27
- 18-Dec-90 — Actual -25 — Shock -21
- 8-Jan-91 — Actual -25 — Shock -18
- 1-Feb-91 — Actual -50 — Shock -25
- 8-Mar-91 — Actual -25 — Shock -16
- 30-Apr-91 — Actual -25 — Shock -17
- 6-Aug-91 — Actual -25 — Shock -15
- 13-Sep-91 — Actual -25 — Shock -5
- 31-Oct-91 — Actual -25 — Shock -5
- 6-Nov-91 — Actual -25 — Shock -12
- 6-Dec-91 — Actual -25 — Shock -9
- 20-Dec-91 — Actual -50 — Shock -28
- 9-Apr-92 — Actual -25 — Shock -24
- 2-Jul-92 — Actual -50 — Shock -36
- 4-Sep-92 — Actual -25 — Shock -22
- 4-Feb-94 — Actual 25 — Shock 12
- 22-Mar-94 — Actual 25 — Shock -3
- 18-Apr-94 — Actual 25 — Shock 10
- 17-May-94 — Actual 50 — Shock 13
- 6-Jul-94 — Actual 0 — Shock -5
- 16-Aug-94 — Actual 50 — Shock 14
- 27-Sep-94 — Actual 0 — Shock -8
- 15-Nov-94 — Actual 75 — Shock 14
- 20-Dec-94 — Actual 0 — Shock 0
- 1-Feb-95 — Actual 50 — Shock 5
- 28-Mar-95 — Actual 0 — Shock 10
- 23-May-95 — Actual 0 — Shock 0
- 6-Jul-95 — Actual -25 — Shock -1
- 22-Aug-95 — Actual 0 — Shock 0
- 26-Sep-95 — Actual 0 — Shock 0
- 15-Nov-95 — Actual 0 — Shock 6
- 19-Dec-95 — Actual -25 — Shock -10
- 31-Jan-96 — Actual -25 — Shock -7
- 26-Mar-96 — Actual 0 — Shock -3
- 21-May-96 — Actual 0 — Shock 0
- 3-Jul-96 — Actual 0 — Shock -5
- 13-Jul-96 — Actual 0 — Shock 0
- 20-Aug-96 — Actual 0 — Shock -4
- 24-Sep-96 — Actual 0 — Shock -13
- 5-Feb-97 — Actual 0 — Shock -3
- 25-Mar-97 — Actual 25 — Shock 3
- 20-May-97 — Actual 0 — Shock -11
- 2-Jul-97 — Actual 0 — Shock -2
- 19-Aug-97 — Actual 0 — Shock -1
- 30-Sep-97 — Actual 0 — Shock 0
- 12-Nov-97 — Actual 0 — Shock -4
- 16-Dec-97 — Actual 0 — Shock -1
- 4-Feb-98 — Actual 0 — Shock 0
- 31-Mar-98 — Actual 0 — Shock 0
- 19-May-98 — Actual 0 — Shock -3
- 1-Jul-98 — Actual 0 — Shock -1
- 18-Aug-98 — Actual 0 — Shock 1
- 29-Sep-98 — Actual -25 — Shock 0
- 15-Oct-98 — Actual -25 — Shock -26
- 17-Nov-98 — Actual -25 — Shock -6
- 22-Dec-98 — Actual 0 — Shock -2
- 3-Feb-99 — Actual 0 — Shock 0
- 30-Mar-99 — Actual 0 — Shock 0
- 30-Jun-99 — Actual 25 — Shock -4
- 24-Aug-99 — Actual 25 — Shock 2
- 5-Oct-99 — Actual 0 — Shock -4
- 16-Nov-99 — Actual 25 — Shock 9
- 21-Dec-99 — Actual 0 — Shock 2
- 2-Feb-00 — Actual 25 — Shock -5
- 21-Mar-00 — Actual 25 — Shock -3
- 28-Jun-00 — Actual 0 — Shock -2
- 26-Jul-00 — Actual 0 — Shock -2
- 22-Aug-00 — Actual 0 — Shock -2
- 3-Oct-00 — Actual 0 — Shock 0
- 22-Aug-00 — Actual 0 — Shock -2
- 3-Oct-00 — Actual 0 — Shock 0
- 22-Aug-00 (duplicate entry in source) — Actual 0 — Shock -2
- 28-Jun-00 — Actual 0 — Shock -2
- 31-Aug-00 — Actual 0 — Shock -2
- 22-Aug-00 (source contains multiple date groupings) — Actual 0 — Shock -2
- 26-Aug-00 — Actual 0 — Shock -2
- 28-Jun-00 (source grouping) — Actual 0 — Shock -2
- 3-Jan-01 — Actual -50 — Shock -38
- 31-Jan-01 — Actual -50 — Shock 0
- 20-Mar-01 — Actual -50 — Shock 6
- 18-Apr-01 — Actual -50 — Shock -43
- 15-May-01 — Actual -50 — Shock -8
- 27-Jun-01 — Actual -25 — Shock 8
- 21-Aug-01 — Actual -25 — Shock 2
- 2-Oct-01 — Actual -50 — Shock -7
- 6-Nov-01 — Actual -50 — Shock -10
- 11-Dec-01 — Actual -25 — Shock 0
- 30-Jan-02 — Actual 0 — Shock 2
- 19-Mar-02 — Actual 0 — Shock -3
- 7-May-02 — Actual 0 — Shock 0
- 13-Aug-02 — Actual 0 — Shock 3
- 24-Sep-02 — Actual 0 — Shock 2
- 6-Nov-02 — Actual -50 — Shock -19
- 10-Dec-02 — Actual 0 — Shock 0
- 25-Jun-03 — Actual -25 — Shock 15
- 18-Mar-03 — Actual 0 — Shock 5
- 6-May-03 — Actual 0 — Shock 4
- 10-Dec-03 — Actual 0 — Shock 0
- 28-Jan-04 — Actual 0 — Shock 0
- 21-Sep-04 — Actual 25 — Shock 2
- 31-Jan-04 — Actual 25 — Shock 2
- 4-May-04 — Actual 0 — Shock -1
- 16-Mar-04 — Actual 0 — Shock 0
- 3-Oct-04 — Actual 0 — Shock -2
- 21-Sep-04 (source grouping) — Actual 25 — Shock 2
- 10-Nov-04 — Actual 25 — Shock 0
- 14-Dec-04 — Actual 25 — Shock 0
- 2-Feb-05 — Actual 25 — Shock 0
- 22-Mar-05 — Actual 25 — Shock 0
- 3-May-05 — Actual 25 — Shock 0
- 31-May-05 — Actual 25 — Shock 0
- 30-Jun-05 — Actual 25 — Shock 0
- 9-Aug-05 — Actual 25 — Shock 0
- 20-Sep-05 — Actual 25 — Shock 1
- 15-Nov-05 — Actual 25 — Shock 0
- 13-Dec-05 — Actual 25 — Shock 0
- 31-Jan-06 — Actual 25 — Shock 0
- 28-Mar-06 — Actual 25 — Shock 0
- 10-May-06 — Actual 25 — Shock -1
- 29-Jun-06 — Actual 25 — Shock -2
- 8-Aug-06 — Actual 0 — Shock -4
- 20-Sep-06 — Actual 25 — Shock 0
- 25-Oct-06 — Actual 0 — Shock 0
- 12-Dec-06 — Actual 0 — Shock 0
- 31-Jan-07 — Actual 0 — Shock 0
- 21-Mar-07 — Actual 0 — Shock 0
- 9-May-07 — Actual 0 — Shock 0
- 28-Jun-07 — Actual 0 — Shock 0
- 7-Aug-07 — Actual 0 — Shock 3
- 18-Sep-07 — Actual -50 — Shock -15
- 31-Oct-07 — Actual -25 — Shock -2
- 11-Dec-07 — Actual -25 — Shock 1
- 30-Jan-08 — Actual -50 — Shock -10
- 18-Mar-08 — Actual -75 — Shock 17
- 30-Apr-08 — Actual -25 — Shock -5
- 25-Jun-08 — Actual 0 — Shock -3
- 5-Aug-08 — Actual 0 — Shock -1
- 16-Sep-08 — Actual 0 — Shock 6
- 29-Oct-08 — Actual -50 — Shock -43
- 16-Dec-08 — Actual -100 — Shock -12

(Note: Appendix Table 2 in the source is presented as a multi-column layout with date/actual/shock triplets across the page; the above lists each triplet as given in the source.)

### Appendix Table 3. Correlation between U.S. and Local Money Market Rates (monthly correlations, 1990–2008)
- Argentina 0.18
- Australia 0.65
- Austria 0.44
- Belgium 0.42
- Brazil 0.18
- Canada 0.77
- Chile 0.60
- China 0.50
- Hong Kong SAR 0.87
- Colombia 0.57
- Czech Republic 0.51
- Denmark 0.39
- Euro Area countries 0.62
- Finland 0.46
- France 0.44
- Germany 0.42
- Greece 0.61
- India 0.25
- Indonesia 0.30
- Ireland 0.34
- Italy 0.51
- Japan 0.42
- Korea, Republic of 0.54
- Malaysia 0.44
- Mexico 0.63
- Netherlands 0.43
- New Zealand 0.70
- Norway 0.34
- Pakistan 0.21
- Peru 0.60
- Philippines 0.48
- Poland 0.50
- Portugal 0.39
- Russian Federation 0.26
- Singapore 0.78
- South Africa 0.54
- Spain 0.50
- Sweden 0.21
- Switzerland 0.50
- Thailand 0.57
- Turkey 0.28
- United Kingdom 0.68

*Source: _wp10278 - Appendix Table 1. List of Countries*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp10278.pdf_