us monetary shocks and global stock prices

J. Finan. Intermediation 21 (2012) 530–547

Contents lists available at SciVerse ScienceDirect

J. Finan. Intermediation

journal homepage: www.elsevier .com/locate/ jfi

US monetary shocks and global stock prices

Luc Laeven a,b,c,⇑, Hui Tong c

a CentER, Tilburg University, P.O. Box 90153, 5000 LE Tilburg, The Netherlandsb CEPR, 77 Bastwick St, London EC1V 3PZ, United Kingdomc International Monetary Fund, Research Department, 700 19th Street, NW, Washington, DC 20431, United States

a r t i c l e i n f o

Article history:Received 5 July 2011Available online 24 February 2012

Keywords:Monetary policyAsset pricesMonetary transmissionFinancial constraints

1042-9573/$ - see front matter � 2012 Internationhttp://dx.doi.org/10.1016/j.jfi.2012.02.002

⇑ Corresponding author at: International MonetarUnited States. Fax: +1 202 623 4740.

E-mail addresses: [email protected] (L. Laeven), h1 A related debate is about whether or not monet

inflation (e.g., Bernanke and Gertler, 2001; Mishkin,

a b s t r a c t

This paper studies how US monetary policy affects global stockprices. We find that global stock prices respond strongly to changesin US interest rates, with stock prices increasing (decreasing)following unexpected monetary loosening (tightening). This impactis more pronounced for sectors that depend on external financing,and for countries whose domestic monetary policy is more alignedwith that of the United States. Using investment data, we presentresults consistent with this effect operating primarily throughchanges in risk premiums as opposed to changes in expectedreturns. These findings suggest that US monetary shocks affectfirms’ stock prices by influencing local interest rates, and offernew evidence that financial frictions play an important role in thetransmission of monetary policy to the real economy.� 2012 International Monetary Fund. Published by Elsevier Inc. All

rights reserved.

1. Introduction

The recent financial crisis has reinvigorated a long-standing debate on the link between monetarypolicy and asset prices. Some have argued that lax US monetary policy fueled an asset price boom bykeeping real interest rates artificially low (e.g., Taylor, 2007), while others do not regard monetary pol-icy as a key contributory factor to the crisis (e.g., Bernanke, 2010).1 By affecting asset prices, monetarypolicy could influence real decisions. Understanding the link between monetary policy and asset prices istherefore critical to understanding the transmission mechanism of monetary policy.

al Monetary Fund. Published by Elsevier Inc. All rights reserved.

y Fund, Research Department, 700 19th Street, NW, Washington, DC 20431,

[email protected] (H. Tong).ary policy should respond to changes in assets prices beyond their impact on2009).

http://dx.doi.org/10.1016/j.jfi.2012.02.002

mailto:[email protected]

mailto:[email protected]

http://dx.doi.org/10.1016/j.jfi.2012.02.002

http://www.sciencedirect.com/science/journal/10429573

http://www.elsevier.com/locate/jfi

L. Laeven, H. Tong / J. Finan. Intermediation 21 (2012) 530–547 531

Existing empirical work commonly finds a negative link, at least in the short run, between mone-tary policy shocks and returns in the stock market, one of the main financial markets.2 However, themagnitude of this effect and the precise channel through which monetary policy affects stock prices re-mains by and large an open question (Boudoukh et al., 1994).

Some studies have employed structural vector autoregressive (VAR) models to disentanglewhether the impact on stock prices operates mostly through changes in expected cash flows, realinterest rates, or risk premiums. For example, Bernanke and Kuttner (2005) analyze how aggregateequity prices react to changes in monetary policy and the economic sources of that reaction usinga structural VAR model. They find that an unanticipated 25 basis point cut in the federal funds ratetarget is associated with about a 1% price increase in broad stock indexes, and that most of this effectoperates through a change in risk premium.3 However, these methods reveal little about the transmis-sion channels of monetary policy (see Bernanke and Gertler, 1995, for an excellent overview of thesechannels).

In theory, monetary policy may influence stock prices by changing future cash flows or by alteringthe rate at which those cash flows are discounted. For example, a fall in interest rates may improvefirms’ growth prospects by allowing interest rate sensitive firms that were unable to afford projectsat higher rates to increase their investment. Alternatively, a fall in interest rates could improve thefirm’s risk profile by lowering the cost of external borrowing, thereby reducing the firm’s risk pre-mium. Both of these channels linking monetary policy to stock prices arise from the presence of finan-cial frictions that give rise to an external finance premium. Either way, theory makes the cross-sectional prediction that the stock price reaction to monetary policy shocks should vary across firmsdepending on their financial dependence.

In this paper, we test this theoretical prediction using data on 20,121 firms in 44 countries byexamining whether US monetary policy shocks disproportionately affect the stock returns of firmsthat are most dependent on external finance.4 Our asymmetric, cross-sectional identification strategyallows for the control of unobserved time-invariant effects that simultaneously affect monetary policyas well as a firm’s stock price, thereby alleviating concerns about endogeneity and simultaneity bias.5

Moreover, our international setting helps sharpen identification because it is rather unlikely that USmonetary policy responds to (technological) developments in foreign countries. Using stock prices asoutcome variable of interest compared to more traditional variables like investment or output has theadditional advantage that stock prices are available at high frequency, allowing us to perform an analysisof short term responses to policy announcements, thereby reducing concerns that results are confoundedby other factors.

Our identification strategy requires exogenous measures of monetary policy shocks and financialdependence. Following Kuttner (2001) and Bernanke and Kuttner (2005), we measure US monetarypolicy shocks using the 1-day change in the price of the 1-month federal funds futures contract onthe day that the FOMC meeting announces a policy rate change. The advantage of this measure is thatit abstracts from monetary policy actions that were already anticipated by the market. We extendtheir analysis to examine how US monetary policy affects stock prices in countries outside of theUnited States. Doing so strengthens our case of treating US monetary policy shocks as exogenous,since US monetary policy is unlikely to be affected in a systematic way by idiosyncratic shocks in othercountries, unless shocks are global and monetary authorities around the globe respond similarly

2 Related work on the link between inflation and stock prices also tends to find a negative link, at least in the short run (Famaand Schwert, 1977; Fama, 1981), and positive stock market responses to disinflation announcements (Henry, 2002).

3 Using a VAR model that incorporates risk aversion and uncertainty, Bekaert et al. (2010) provide empirical evidence of a linkbetween monetary policy and risk aversion in financial markets. They find that lax monetary policy decreases risk aversion with alag of about five months, with the effect lasting for about 2 years.

4 A large empirical literature has tried to assess the importance of a bank credit channel of monetary policy using cross-sectionalvariation across banks (e.g., Bernanke and Blinder, 1992; Kashyap et al., 1993; Kashyap and Stein, 2000; Jiménez et al., 2009) andtheir subsidiaries (e.g., Peek and Rosengren, 2000; Campello, 2002; Ashcraft and Campello, 2007; Cetorelli and Goldberg,forthcoming). The difference between our paper and this literature is that we focus on asset prices rather than the quantity orquality of credit.

5 A similar cross-sectional approach has been taken by Kashyap and Stein (1994) to examine the asymmetric impact ofmonetary policy on the lending behavior of different types of banks.

532 L. Laeven, H. Tong / J. Finan. Intermediation 21 (2012) 530–547

to such shocks.6 By taking an international perspective on the transmission channel of US monetary pol-icy, this paper sheds light on the role of US monetary policy in influencing global asset prices andasset allocation.

We measure an industry’s financial dependence also using US data, following the influential workby Rajan and Zingales (1998), to gauge the intrinsic demand for external finance in the absence offinancial constraints. This approach relies on the assumption that large US listed firms face minimalfinancial constraints given the depth of US financial markets and that the ranking of financial depen-dence across sectors in the US is preserved in other countries.7

We find strong evidence of a negative response of stock prices to US monetary shocks, with USmonetary policy loosening (tightening) being associated with an increase (decrease) in stock pricesin other countries, consistent with earlier work based on US stock prices (e.g., Thorbecke, 1997;Ehrmann and Fratzscher, 2004; Bernanke and Kuttner, 2005).8 Moreover, this impact is particularlypronounced for firms with a relatively high intrinsic dependence on external finance. For example, anunexpected policy rate decrease of 5 basis points (equal to its interquartile range) is associated with astock price response that is 6 basis points greater for firms whose financial dependence is at the 75thpercentile (the Construction machinery industry) relative to firms whose financial dependence is atthe 25th percentile (the Beverages industry). This is a significant effect compared to the average stockmarket return around FOMC dates of 19 basis points.

One concern is that these results may capture an international transmission of real shocks in whichUS monetary policy changes affect foreign firms (for example, their profitability) through their impacton US real variables, such as expectations about US real GDP growth. To lend additional support to ourinterpretation of a monetary policy channel we show that the impact of monetary shocks on stockprices is stronger in countries with pegged currencies (as opposed to flexible exchange rate regimes)and in countries whose domestic monetary policy is more aligned with that of the United States (asmeasured by the correlation of domestic and US interest rates). In both cases, US monetary policyis likely to affect local monetary policy and the transmission is likely to be through monetary shocksrather than through US real shocks. This suggests that US monetary shocks affect firms’ stock pricesthrough local interest rates.

To further distinguish financial from real explanations, we consider the role of local business cycles.Reassuringly, our main result on the interaction between financial dependence and the monetary pol-icy shock is little affected when controlling for the effect of local recessions. To mitigate concerns thatthe international transmission of shocks is driven by globalization of firms, we show that the resultsare robust to excluding firms with assets abroad and to excluding cross-listed firms.

To further reduce concerns that alternative channels may be driving the results, all regressions in-clude country-time and firm fixed effects (unless otherwise noted), thereby effectively controlling foralternative channels that affect firms differently depending on their country of location and the timeperiod in question. We report standard errors that are corrected for two-way clustering by sector andtime, although results are robust to alternative dimensions of clustering of standard errors, includingthree-way clustering by country, sector and time.

Taken together, these results suggest that financial frictions play an important role in the trans-mission of monetary policy, and that US monetary policy influences global asset prices primarilythrough local interest rates and by affecting disproportionally the stock prices of financially dependentfirms.

6 Cetorelli and Goldberg (forthcoming) also study how US monetary policy shocks are transmitted abroad. Rather than analyzingtheir impact on stock prices of non-financial firms, they study how US monetary policy affects lending activity abroad by foreignsubsidiaries of US banks. They find that the globalization of banking has weakened the lending channel of monetary policydomestically but has made lending abroad more sensitive to US monetary policy shocks. Wongswan (2006) uses high-frequencydata to study the effect of US monetary policy shocks on the volatility and trade volume of Korean and Thai equity markets, butdoes not investigate transmission channels.

7 Following Rajan and Zingales (1998), who use this approach to study the impact of financial development on economic growth,this approach has been applied, among others, to study the role of business cycles (Braun and Larrain, 2005), demand for workingcapital (Raddatz, 2006), and financial crises (Kroszner et al., 2007) in influencing the link between finance and growth.

8 Ehrmann and Fratzscher (2004) also test whether the strength of this reaction is dependent on proxies for the credit channelfor US firms.


Empirical research on the link between monetary policy shocks and stock prices has generally notconsidered the role of financial constraints. Bernanke and Kuttner (2005) and Boudoukh et al. (1994)analyze the differential impact of monetary policy shocks on stock prices across broad classes ofindustries but do not explicitly consider the role of financial constraints, while Kashyap et al. (1994)do consider financial constraints to show that firm inventory investment by liquidity constrained firmsis significantly adversely affected during periods of tight monetary policy but they do not analyze itsimpact on stock prices. Similarly, Gertler and Gilchrist (1994) show that the investment of small firms,their proxy for the importance of financial frictions, responds more strongly to monetary policy thanthat of large firms. Finally, research on stock prices and financial constraints has generally not consid-ered the role of monetary policy (e.g., Baker et al., 2003; Gomes et al., 2006; Livdan et al., 2009). Anexception is Lamont et al. (2001) who find little role for monetary policy but they use traditionalmeasures of monetary policy that do not disentangle monetary shocks from market expectations.9

More recently, Ammer et al. (2010) study how US monetary shocks affect the stock prices of foreign stockscross-listed in the US market. We study a much larger sample of foreign stocks that also includes firmsthat are not cross-listed in the US market.10

The paper proceeds as follows. Section 2 presents our empirical strategy, construction of key vari-ables, and sources of data. Section 3 discusses the main empirical results and a slew of robustnesschecks and extensions. Section 4 offers concluding remarks.

2. Methodology and data

2.1. Methodology

Our basic empirical strategy is to test whether an exogenous monetary shock in the US has an im-pact on the stock return of firms in other countries, and whether this effect is more pronounced forfirms that are more financially dependent. We use the approach in Bernanke and Kuttner (2005) toidentify unexpected policy rate changes, and build on their work by extending the analysis to othercountries and by considering asymmetric responses across firms depending on their degree of finan-cial dependence. This allows us to deal more effectively with concerns about endogeneity and simul-taneity, and discern more precisely one of the channels through which monetary policy affects stockprices.

Our analysis starts by confirming the common finding in the literature that stock returns are neg-atively associated with innovations in monetary policy. We do this by showing that stock prices re-spond negatively to unanticipated changes in the US federal funds rate following meetings of theUS Federal Open Market Committee (FOMC). To be precise, we estimate the following equation:

9 In rconstra

10 Haustock pon firm

Stock Returni;j;k;t ¼ b Monetary Shockt þ c Controli;j;k;t þ ei;j;k;t ð1Þ

where i stands for company, j for country, k for sector, and t for time. Note that this is a panel regres-sion. We start by assuming the same b for all sectors and countries in order to estimate an averageeffect, but will subsequently allow for variations across sectors, countries, and time. We further in-clude firm and country-time specific fixed effects to control for unobserved firm and country-timespecific factors, and cluster standard errors at both the FOMC date level and the 3-digit SIC sector level.

To investigate how an industry’s financial dependence affects the impact of the US monetary policyshock, we now consider the interaction between the monetary policy shock and an industry’s depen-dence on external finance. In other words

Stock Returni;j;k;t ¼ b1 Monetary Shockt þ b2 Financial Dependencek �Monetary Shockt

þ c Controli;j;k;t þ ei;j;k;t ð2Þ

elated work, Campello and Chen (2010) consider the impact of macroeconomic shocks on stock returns of financiallyined firms, but they do not consider the role of monetary policy explicitly.sman and Wongswan (2011) find a role of exchange rate regimes in the spillover of FOMC announcement effects on global

rices, while Ehrmann and Fratzscher (2009) find a role for the degree of financial integration but neither assesses the impact-level stock returns and differential effects based on external financial dependence.


where Financial Dependencek measures the external financing needs for capital expenditure for firms ina given industry following Rajan and Zingales (1998). The slope coefficient, b2, then captures the ex-tent to which the effect of the monetary policy shock depends on an industry’s dependence on exter-nal financing.

Finally, we examine if the impact of US monetary shocks on stock prices varies across countries byincluding additional interaction terms between country characteristics (such as economic and finan-cial development) and the monetary shock variable. In other words, we extend Eq. (2) as follows:

11 Alte(2004),and is tright frpublicly

12 Thisystema

Stock Returni;j;k;t ¼ b1 Monetary Shockt þ b2 Financial Dependencek �Monetary Shockt

þ b3Country Traitjt �Monetary Shockt þ c Controli;j;k;t þ ei;j;k;t ð3Þ

where the slope coefficient b3 then captures the extent to which the effect of the monetary policyshock depends on a particular country trait.

2.2. Data and variable definitions

2.2.1. Stock pricesTo construct our dependent variable, we collect data on stock prices of 20,121 firms in 44 countries

over the period 1990–2008. Appendix A shows the complete list of countries. Stock price data are re-trieved from Datastream, and are adjusted for dividends and capital actions such as stock splits andreverse splits. We consider 2-day stock price responses to monetary policy shocks following FOMCmeetings. Specifically, we compute the stock return as the log difference in the closing price of thestock over the period t � 1 and t + 1, where t is the day of the FOMC meeting. The reason for usinga 2-day window rather than a 1-day window of stock returns is due to time zone differences betweenstock markets in the US and other countries. Our sample includes a total of 142 FOMC meetings. Toreduce the impact of extreme values, we drop 2-day stock returns with a value of above 50% or below�50%, which covers 0.1% of the sample. As a robustness check, we also winsorize the sample at its topand bottom 1% level, and our results are unaltered.

2.2.2. Monetary policy shockOur measure of monetary shocks at US FOMC meetings follows the approach in Kuttner (2001) and

Bernanke and Kuttner (2005).11 They propose to use the change in the price of federal funds futures con-tracts relative to the day prior to the policy action as a measure of unexpected policy rate changes.12 For aFOMC meeting on day d of month m, the monetary shock is the change in the rate implied by the current-month futures contract. However, because the contract’s settlement price is based on the monthly aver-age federal funds rate, the change in the implied futures rate then should be scaled up by a factor relatedto the number of days in the month affected by the change, or:

Shockd ¼D

D� dðfd � fd�1Þ ð4Þ

where Shock is the unexpected target rate change, fd is the current-month futures rate at day d, and Dis the number of days in the month. The expected rate change then is the actual change minus theshock.

We extend the data coverage of monetary shocks from year 2002 to year 2008. Moreover, whileBernanke and Kuttner (2005) exclude FOMC dates when there is no policy rate change, we includethese dates as well as control groups. Appendix B lists the exact dates of FOMC meetings, the actualchanges of federal funds rate, and the unexpected component of the change.

rnative measures of monetary policy include those developed by Bernanke and Mihov (1998) and Romer and Romeramong others. The measure of monetary policy by Bernanke and Mihov is computed using VAR models on monthly dataherefore not applicable to our empirical approach for which daily data are needed. The Romer and Romer measure has theequency but uses information from the Fed’s green books that include Fed staff economic forecasts that are not made

available to the market until five years after the FOMC meeting and may therefore not be fully incorporated in stock prices.s approach assumes that risk premia that could be embedded in prices on federal funds futures do not changetically within a 1 day period (Piazzesi and Swanson, 2008).


2.2.3. Financial dependence indexAs measure of an industry’s intrinsic dependence on external finance, we use the financial depen-

dence measure proposed by Rajan and Zingales (1998). They compute an industry’s dependence onexternal finance as:

Table 1Summalog diffechangecurrentand thebased o(2003)total of

Key

StocActuExpeShocFinaTang

Financial dependence ¼ Capital expenditures� Cash flowCapital expenditures

; ð5Þ

where cash flow = cash flow from operations + decreases in inventories + decreases in receiv-ables + increases in payables. The index is computed using data on publicly listed US firms, whichare judged to be least likely to suffer from financing constraints relative to generally smaller, non-listed US firms and firms in other countries. Conceptually, the Rajan and Zingales index aims to iden-tify sectors that are naturally more dependent on external financing for their business operation.While the original Rajan and Zingales (1998) paper covers only 40 (mainly SIC 2-digit) sectors, werecompute their measure using data for the period 1990–2006 to expand the coverage to around150 SIC 3-digit sectors. We drop firms active in the utilities industry (SIC 4), wholesale and retailindustry (SIC 5), financial industry (SIC 6), and public administration (SIC 9) because these firms aresubject to strict regulation or because their financing needs are not comparable with those of otherindustries.

To calculate the demand for external financing of US firms, we take the following steps. We firstsort every firm in the Compustat USA files based on their 3-digit SIC sectoral classification and thencalculate the ratio of dependence on external finance for each firm by aggregating cash flows andexpenditures as in Rajan and Zingales over the period 1990–2006. We then calculate the financialdependence index as the sector-level median value of these firm ratios for each SIC 3-digit sector thatcontains at least 5 firm observations.

2.3. Descriptive statistics

Table 1 reports summary statistics of the key dependent and explanatory variables. The actualchange in the federal funds rate announced following FOMC meetings ranges from a rate cut of 100basis points to a rate increase of 75 basis points. Unexpected rate shocks vary from �43 basis pointsto +24 basis points, indicating that rate surprises on average were on the downside.

Financial dependence ranges from a low of �2.4 for the Manifold Business Forms industry (anindustry that has seen a decline in activity since the 1990s due to the growing use of computers) toa high of 1.4 for the Photographic Equipment and Supplies industry (an industry that has gone digitaland hence seen large capital investment).

ry statistics. This table reports summary statistics for the main variables in our analysis of non-US firms. Stock return is therence in the closing price over the period t � 1 and t + 1, where t is the FOMC meeting date. The actual change is the actualin federal funds rate announced at FOMC meetings. The shock is the change in the federal funds rate implied by the

-month futures contract from Bernanke and Kuttner (2005). The expected change is the change between the actual changeshock in federal funds rate. Financial dependence is an industry’s intrinsic dependence on external finance for investmentn Rajan and Zingales (1998). Tangibility is an industry’s intrinsic ratio of tangible assets over total assets based on Braunand Claessens and Laeven (2003). The sample consists of 20,121 firms across 44 countries over the period 1990–2008, for a925,306 firm-time observations.

variables Obs Mean Median St. dev. P25 P75 Min Max

k return (in%) 925,306 0.19 0 5.38 �1.79 2.00 �50 50al change (in bp) 140 �4.82 0 26.29 �25 0 �100 75cted change (in bp) 140 �1.31 0 22.19 �8.5 4.5 �92 61k (in bp) 140 �3.51 0 10.63 �5 0 �43 24ncial dependence 150 �0.06 0.04 0.7 �0.37 0.33 �2.4 1.4ibility 150 0.30 0.25 0.17 0.17 0.39 0.06 0.78


3. Empirical results

3.1. Baseline results

In what follows, we use US monetary shocks as our source of exogenous variation in monetary pol-icy for foreign firms. This lends additional credibility to using our measure of monetary shocks as anexogenous variable of monetary policy given that US monetary policy is unlikely to respond in a sys-tematic way to idiosyncratic economic factors in other countries, though it may respond to economicfactors in the United States.

The baseline results for foreign firms are presented in Table 2. We include firm fixed effectsthroughout, and cluster standard errors at both sector and FOMC meeting date levels. In Column 1,we examine the impact of actual federal funds rate changes on stock returns. We obtain a negativecoefficient on the actual federal funds rate variable but this coefficient is not significantly differentfrom zero.

In Column 2, we further decompose the change of the federal funds rate into its expected and unex-pected components, following the method proposed by Bernanke and Kuttner (2005). We find that theunexpected component has a significantly negative coefficient. Based on the estimated coefficient of0.04, a 25 basis point increase in US rates would reduce global stock prices by about 1%. Also, based onthe coefficient of 0.04, a one standard deviation increase in the monetary shock would reduce stockprices by 0.4%, which would explain 8% of the standard deviation of stock returns. The expected ratechange component enters with a positive coefficient, but it is less statistically significant and its eco-nomic impact is only about 10% of the impact for the unexpected rate shock.

In Column 3, we focus on the unexpected rate shocks only. We find that the unexpected rate shockvariable has a coefficient that is similar to that in Column 2. This indicates that unexpected and ex-pected rate shocks are orthogonal to each other, and the exclusion of the expected rate shock variable,which should be incorporated in stock prices, does not alter the results we find for the unexpected rateshock variable.

In Column 4, we examine the asymmetric impact of monetary shocks on stock prices based on sec-tor-level dependence on external finance for capital expenditure, using the Rajan and Zingales (1998)

Table 2The effect of monetary policy shocks on stock returns. Dependent variable is stock return measured as the log difference in theclosing price over the period t � 1 and t + 1, where t is the FOMC meeting date. The actual change of federal funds rate (FFR) is thechange announced at FOMC meetings. The unexpected change of FFR (shock) is the change in the FFR implied by the current-month futures contract from Bernanke and Kuttner (2005). Financial dependence is an industry’s intrinsic dependence on externalfinance for investment based on Rajan and Zingales (1998). Standard errors are clustered by sector and time. Robust standarderrors in brackets.

Actualchange(1)

Components(2)

Shock(3)

Financialdependence(4)

Timeeffects(5)

Country-timeeffects(6)

Actual change �0.003[0.004]

Expected change 0.005*

[0.003]

Shock �0.043** �0.041** �0.038**

[0.017] [0.017] [0.016]

Shock � financialdependence

�0.017*** �0.014*** �0.011***

[0.003] [0.002] [0.002]

Firm fixed effects Y Y Y Y Y YTime fixed effects N N N N Y YObservations 925,306 925,306 925,306 925,306 925,306 925,306R-squared 0.025 0.031 0.031 0.031 0.063 0.139

*** p < 0.01.** p < 0.05.* p < 0.1.


measure of financial dependence. We do this by including the interaction term of financial depen-dence and the unexpected shock variable. We do not include the financial dependence variable by it-self in the regression specification, as it is fully absorbed by the firm dummies that we include. We findthat the impact of interest rate shocks is statistically significantly higher for sectors that depend moreon external finance. Based on the estimated coefficients in Column 4, an unexpected policy rate de-crease of 5 basis points (equal to its interquartile range) is associated with a stock price response thatis 6 basis points greater for firms whose financial dependence is at the 75th percentile (the Construc-tion machinery industry) relative to firms whose financial dependence is at the 25th percentile (theBeverages industry). This is a significant effect compared to the average stock market return aroundFOMC dates of 19 basis points.13

In Column 5, we include dummies for each FOMC date to control for other contemporaneous timeeffects. The Shock variable is fully captured by these time dummies and is therefore dropped from thisregression. The interaction of our variable of interest, Shock � Financial dependence, still has a signifi-cant coefficient of �0.014, comparable to the results without FOMC date effects.14

Finally, in Column 6 we include dummies for country-time pairs. Not surprisingly, the coefficient ofShock � Financial dependence reduces somewhat to �0.011, due to the correlation between this vari-able and the country-time dummies. Importantly, however, it remains significant at the 1% level.15

Our results thus far would be biased if monetary policy were to respond contemporaneously to thestock market. However, Bernanke and Kuttner (2005) do not find any evidence for such a systematicreaction. Moreover, if the FOMC did respond to large changes in equity prices, such reverse causalitywould tend to reduce the size of the estimated response to the monetary shock and would thereforebias our results toward finding no effect.

Additionally, results would be biased if monetary policy and stock prices were jointly determinedand responded simultaneously to new information. For instance, bad news about the health of theeconomy could lead to both negative stock price reactions and a loosening of monetary policy. Thiswould lead to a downward bias in the size of the estimated response to the monetary shock, biasingour results toward finding no effect. This mitigates concerns that our results are driven by reverse cau-sality or simultaneity bias.

3.2. Robustness checks

3.2.1. Additional firm and sector characteristicsTo further address concerns about simultaneity bias and omitted variables, we now control in Table

3 for the potential effect of channels other than those arising from financial constraints. Of particularinterest is the demand channel. For example, when there is a negative monetary policy shock associ-ated with a contraction in demand, sectors with superior growth opportunities may suffer relativelymore. We therefore consider the possibility that differences in growth opportunities may influence theresult. In Column 1, we control for this alternative channel using an industry’s lagged sales growthbased on US data as proxy for growth opportunities, similar to Fisman and Love (2007). Specifically,we include this measure of global growth opportunities and its interaction with financial dependenceinto our regression model. This interaction enters with a negative coefficient, albeit insignificant, sug-gesting that monetary shocks may also weakly affect stock prices through the demand channel.

13 We also explore the case without firm fixed effects. In a specification similar to Column 4 but without firm fixed effects, wefind Shock � financial dependence has a coefficient of �0.017 that is statistically significant at the 1% level, similar to thespecification with firm fixed effects.

14 The results are consistent across alternative levels of clustering. In unreported tables, we use the following levels of clustering:country level, sector level, country and date levels (two-way), country and sector levels (two-way), sector and date levels (two-way), and country, sector and date levels (three-way). Our main variable of interest, Shock � Financial dependence, remainssignificant at the 1% level throughout these alternative clustering specifications.

15 We also examine how US monetary shocks affect US firms. In a regression similar to that in Column 6 of Table 2, the coefficientof Shock � Financial dependence is �0.028 with a standard error of 0.01, for a total of 463,158 US firm-time observations. Thisimpact is about three times that for international firms, suggesting that US monetary policy has more direct effects on U.S. firms.While these results serve as a useful benchmark, identification is complicated by the possibility that U.S. monetary policy mayrespond to the performance of US firms.

Table 3Robustness checks. Dependent variable in Columns 1–4 is stock return measured as the log difference in the closing price over theperiod t � 1 and t + 1, where t is the FOMC meeting date. In Column 5, dependent variable is the abnormal stock return estimatedusing a market model over the period t � 1 and t + 1, where beta varies annually based on the correlation of weekly firm-levelstock returns and the local market returns in the year of the FOMC date. In Column 6, dependent variable is stock return measuredas the log difference in the closing price over the period t � 1 and t, with time-zone adjustment. The unexpected change of FFR(shock) is the change in the FFR implied by the current-month futures contract from Bernanke and Kuttner (2005). Financialdependence is industry’s intrinsic dependence on external finance for investment based on Rajan and Zingales (1998). Globalgrowth opportunity is the one-period lagged sectoral growth rate in sales for US firms, similar to Fisman and Love (2007). Thedefinition of durable goods follows Braun and Larrain (2005) and the classification in the Bureau of Economic Analysis’s IndustryAccounts. Tangibility is tangible assets over total assets, similar to Braun (2003). Column 4 focuses on firms whose foreign assetsare less than 0.01% of total assets and have never been cross-listed in the US during the sample period. Standard errors areclustered by sector and time. Robust standard errors in brackets.

Growthopportunity

Durablegoods

Tangibility No foreign assetsor cross-listing

Abnormalreturn

One-dayreturn

(1) (2) (3) (4) (5) (6)

Shock � financial dependence �0.011*** �0.011*** �0.011*** �0.011*** �0.004** �0.01***

[0.002] [0.003] [0.002] [0.002] [0.002] [0.001]

Shock � growth opportunity �0.033[0.024]

Shock � durable goods �0.001[0.002]

Shock � tangibility �0.007[0.030]

Growth opportunity 0.100[0.150]

Firm fixed effects Y Y Y Y Y YCountry-time fixed effects Y Y Y Y Y YObservations 923,107 925,306 925,306 705,304 925,306 925,306R-squared 0.14 0.14 0.14 0.14 0.10 0.11

*** p < 0.01.** p < 0.05.� p < 0.1.


Importantly, though, our main results on the Shock � Financial dependence interaction variable are notaltered.

In Column 2, we add a durable goods variable as another proxy for demand channel. The classifi-cation of durable/non-durable goods follows Braun and Larrain (2005) and Raddatz (2006), which isbased on the classification of durable goods presented in the Bureau of Economic Analysis’s IndustryAccounts. Durable goods are assigned as 1, nondurable goods as 0, and semi-durable goods as 0.5.16

The interaction term of durable goods and monetary shock does not turn out to be significant. Moreimportantly, the inclusion of durable goods does not affect our results for financial dependence.

In Column 3, we further analyze whether the effect of monetary shock varies with the asset tangi-bility of a sector, as tangible assets tend to be used as collateral for borrowing and thus tend to boost afirm’s capacity to borrow external funds. We follow Braun (2003) and Claessens and Laeven (2003) toconstruct the sector-level tangibility index based on the Compustat data for US firms. We take the fol-lowing steps. We first calculate the tangibility of each US firm annually from 1990 to 2006. We dividenet property, plant, and equipment (Compustat item 33) by total assets (Compustat item 6) to con-struct the firm-level tangibility. We then sort every firm based on their 3-digit SIC sectoral classificationand calculate the sector tangibility index as the sector-level median value of these firm-year tangibilityratios for each SIC 3-digit sector. We then apply the sector tangibility index to our sample of countries,and include the interaction term of this tangibility index and the monetary policy shock variable in theregression in Column 3.17 We find that Shock � Tangibility has a negative but insignificant coefficient. The

16 The semi-durable industries are clothing, footwear, and printing.17 We do not include the tangibility index itself as it is already captured by firm-level fixed effects.


negative sign is intuitive in that an unexpected monetary policy tightening may lower the value of collat-eral, thereby reducing a firm’s capacity to borrow, and consequently its stock price. Reassuringly,Shock � Financial Dependence has a negative coefficient that remains significant at the 1% level.

One competing explanation for the international effect we find is that firms themselves are glob-alized. For example, some multinational firms may have subsidiaries operating in the US, while somefirms are listed in both domestic and US stock markets. Hence, in Column 4, we exclude firms withforeign assets that exceed 0.1% of their total assets.18 Meanwhile, we also exclude firms that have everbeen cross-listed in the US markets. We find that Shock � Financial dependence obtains the same coeffi-cient and remains significant at the 1% level.

3.2.2. Abnormal returnsIn the literature, abnormal stock returns have also been examined to study the impact of

macroeconomic shocks (e.g., Mackinlay, 1997). To compute the abnormal return, one has to firstcalculate the normal stock return. There are two common models for the normal return: the constantmean return model and the market model. The constant mean return model assumes that the meanreturn of a given security is constant through time. The market model assumes a stable linearrelation between the market return and the security return. Our estimations have used the constantmean return so far, as we include firm-level fixed effects in the regressions. As a robustness check, wenow also employ the market model to construct the abnormal return as our dependent variable asfollows:

18 We19 We

models20 Not

compar

Abnormal returni;t ¼ Stock returni;t � Alphai;t � Betai;t �Market returnj;t ð6Þ
We first construct each firm’s beta annually based on the correlation of weekly firm-level stock re-
turns and the local market returns.19 We then construct each firm’s alpha in a given year as the annualaverage of the firm’s weekly average return minus the beta multiplied by the annual average market re-turn. We winsorize the generated abnormal returns around each FOMC date at the 1% level. The regressionresults when using these abnormal returns and firm fixed effects are presented in Column 5 of Table 3. Wefind that Shock � Financial dependence continues to obtain a negative coefficient significant at the 5% level.

3.2.3. Trading times of stock marketsSo far we have examined the stock return between t � 1 and t + 1. As a robustness check, we also

examine the stock returns between t � 1 and t, where t is the FOMC meeting date. One complicationwhen using 1-day stock returns is the need to control for the time-zone difference with respect to theclosing time of each stock market relative to the FOMC announcement time. We consider three majortime zones: the Americas, Europe, and Asia. For America, t will be the US FOMC meeting date (in USEastern Time). Adjusting returns for time-zone differences is more complicated for European marketsdue to the fluctuation in FOMC announcement times. Since 1994, FOMC announcements tend to occurat 2:15 pm. Before the year 1994, they varied from anywhere between 8:30 am to 3:30 pm (see theAppendix in Gürkaynak et al. (2005) for more details). Hence, we compare the FOMC announcementtime with the market closing time in the European market on a case-by-case basis to adjust t. For mar-kets in Asia, t is set equal to 1-day after the FOMC meeting date (in US Eastern Time) because Asianmarkets are generally closed when the FOMC meeting decision is announced. The results are pre-sented in Column 6 of Table 3. We find that the results for the time-zone adjusted 1-day returnsare qualitatively similar to those obtained when using 2-day returns.

3.2.4. FOMC meetings without rate changeOur results thus far are based on a sample that includes observations from FOMC meetings on

which no change in the federal funds rate took place.20 We want to make sure that the inclusion of

set missing values for foreign assets to zero. Excluding these firms as well does not alter our findings.use the domestic beta rather than a beta based on a world factor model because Griffin (2002) finds that domestic factorperform better in explaining time-series variations in returns and have lower pricing errors than the world factor model.e that Bernanke and Kuttner (2005) dropped such cases from their analysis. This robustness check therefore also enhancesison of their results with ours.


dates on which no rate change took place does not confound our results. We therefore drop cases wherethere is no change in the federal funds rate at the FOMC meetings in a robustness check. In unreportedregressions, we find that the results when excluding observations without a change in policy rate arequite similar to the baseline case in Table 2.

3.2.5. Manufacturing sectors onlyIn the original paper by Rajan and Zingales (1998), financial dependence was computed for man-

ufacturing industries only due to data limitations. In addition, the index of financial dependence theydevelop is less applicable to industries without significant capital expenditures and to heavily regu-lated sectors. We already excluded the utility sector, trade sector, financial sector, and governmentsector from the sample to accommodate this. Next, we rerun our main regression specification onthe subset of manufacturing firms in our sample. In unreported regressions, we find similar resultsfor the impact of US monetary shocks for manufacturing firms. Our main variable of interest,Shock � Financial dependence, remains statistically significant at conventional levels.

3.3. Country factors

In Table 4, we examine whether and how the impact of US monetary shocks varies acrosscountries. Throughout the specifications, we include firm fixed effects and FOMC meeting fixedeffects.

In Column 1, we examine the impact of financial development by adding a proxy for domesticfinancial development and its interaction with the monetary policy shock. A large literature starting

Table 4Additional country factors. Dependent variable is stock return measured as the log difference in the closing price over the periodt � 1 and t + 1, where t is the FOMC meeting date. Financial development is a country’s domestic private credit over GDP. Per capitaincome is GDP per capita in constant US dollar term. Local recession is a dummy following Braun and Larrain (2005). We measurepolicy synchronization by the correlation of monthly money market rates between the US and the individual countries over theperiod from 1990 to 2008. Standard errors are clustered by sector and time. Robust standard errors in brackets.

Financial development GDP per capita Local recession Policy synchronization

Shock � financial dependence �0.014*** �0.013*** �0.014*** �0.013***

[0.001] [0.002] [0.002] [0.001]

Shock � financial development �0.005[0.010]

Shock � per capita income �0.008[0.010]

Shock � local recession �0.01[0.012]

Shock � policy synchronization �0.052***

[0.012]

Financial development �0.110[0.180]

Per capita income �0.360[0.440]

Local recession 0.013[0.13]

Firm fixed effects Y Y Y YTime fixed effects Y Y Y YObservations 875,140 925,306 925,306 915,823R-squared 0.06 0.06 0.06 0.06

*** p < 0.01.�� p < 0.05.�p < 0.1.


with King and Levine (1993) has shown that financial development boosts economic growth byrelaxing financial constraints. Following this literature, we measure domestic financial developmentby domestic credit to private sector over GDP. The interaction term of domestic financial developmentand the shock variable does not turn out to be significant.

In Column 2, we control for the impact of economic development of the country, as proxied by logper capita GDP. We include log per capita GDP and its interaction with the US monetary shock vari-able. Neither of these additional variables enters significantly. More importantly, the interaction ofShock � Financial dependence still keeps its magnitude and significance.

In Column 3, we further consider the role of local recessions as defined in Braun and Larrain(2005).21 We include the interaction of US monetary shock with a local recession dummy that takes avalue of one if the country is in a recession and zero otherwise. This interaction term has a negative coef-ficient, but it is not statistically different from zero. Reassuringly, Shock � Financial dependence keeps itsmagnitude and significance.

In Column 4, we examine potential asymmetric impact across countries based on the degree of pol-icy rate synchronization. We expect that our effect is more pronounced in countries whose monetarypolicy is more closely aligned with US monetary policy in general. As measure of synchronization ofmonetary policy, we use the correlation of monthly money market rates between the US and the indi-vidual countries over the period from 1990 to 2008 (for Euro countries we substitute the country’smarket rate with the Euro money market rate after the country joins the Euro). (Appendix C liststhe estimated correlation coefficient). Not surprisingly, we find that the asymmetric impact of mon-etary shocks on the return of firms is more significant in countries with a high synchronization ofmonetary policy with the United States. This suggests that US monetary shocks affect firms’ stockprices through local interest rates.

In Table 5, we further examine the role of US monetary shocks and the relevance of the monetarypolicy transmission channel by performing sample splits. Specifically, we test whether US monetaryshocks have stronger effects in countries whose monetary policy is more aligned with that of the USand in countries with less flexible exchange rates. US monetary shocks are expected to have a largereffect on local interest rates in countries whose monetary policy closely follows that of the US and incountries whose currency is pegged to the dollar (and where therefore interest rates have to accom-modate exchange rate policy). If we therefore find that monetary shocks affect the stock prices offinancially dependent firms more in countries whose monetary policy is more aligned with thatof the US and in countries with less flexible exchange rates, this would lend support to our resultsbeing driven by the transmission of monetary policy as opposed to the international transmission ofreal US shocks. In Columns 1 and 2, we split the sample between non-flexible versus flexible ex-change rate regimes.22 For non-flexible regimes, we find that Shock � Financial dependence obtains acoefficient of �0.016, significant at the 1% level. On the contrary, for flexible regimes, the coefficienton Shock � Financial dependence is only �0.008. In Columns 3 and 4, we split the countries betweenhigh versus low synchronization of monetary policy.23 We find that the coefficient of Shock � Financialdependence is twice as large in countries with a relatively high degree of policy synchronization com-pared to countries with relatively low synchronization. These results lend support to the relevance ofthe monetary policy transmission channel whereby US monetary shocks affect local policy rates, espe-cially in countries with less flexible exchange rates where interest rate policy has to accommodate ex-change rate policy.

21 The definition of recession follows a peak-to-trough criterion as in Braun and Larrain (2005). A trough occurs when cyclicalGDP is at least one standard deviation below zero. From the trough, we then go backwards in time until a local peak, which is a yearwhen cyclical GDP is above both the previous and posterior years. The recession dummy then takes a value of one from the yearafter the peak to the trough.

22 The classification of exchange rate regimes is derived from Reinhart and Rogoff (2004). Our definition of flexible exchange rateregimes includes managed floating regimes, freely floating regimes, free falling regimes, and Euro area countries after they adoptedthe euro.

23 A country has a high (low) degree of policy synchronization if the correlation between US and local money market rates isabove (below) 0.5.

Table 5Monetary shocks, exchange rate regimes, and synchronization of monetary policy. Dependent variable is stock return measured asthe log difference in the closing price over the period t � 1 and t + 1, where t is the FOMC meeting date. The unexpected change ofFFR (shock) is the change in the FFR implied by the current-month futures contract from Bernanke and Kuttner (2005). Financialdependence is industry’s intrinsic dependence on external finance for investment based on Rajan and Zingales (1998). Columns 1and 2 are for countries with non-flexible and flexible exchange rate regimes respectively, per the classification in Reinhart andRogoff (2004). We measure policy synchronization by the correlation of monthly money market rates between the US and theindividual countries over the period from 1990 to 2008. A country has high (low) policy synchronization if the correlation is above(below) 0.5. Columns 3 and 4 are for high and low synchronization countries respectively. Standard errors are clustered by sectorand time. Robust standard errors in brackets.

Non-flexibleregimes

Flexibleregimes

High PolicySynchronization

Low PolicySynchronization

Shock � financial dependence �0.016*** �0.008** �0.015*** �0.006***

[0.003] [0.003] [0.002] [0.002]

Firm fixed effects Y Y Y YCountry-time fixed effects Y Y Y YObservations 404,523 518,772 473,881 441,942R-squared 0.15 0.14 0.11 0.18

*** p < 0.01.** p < 0.05.� p < 0.1.


Overall, we find a strong and robust asymmetric relationship between the stock responses of finan-cially dependent firms and monetary policy shocks, even with the various controls of other character-istics at the firm, sector, or country level.

3.4. What explains the link between monetary policy shocks and stock prices?

Thus far, we have shown that monetary policy disproportionately influences the stock prices offirms that depend on external financing but we have not shown whether this channel operatesthrough changing future cash flows or through altering the rate at which those cash flows are dis-counted. The latter can arise from changes in the real interest rate or changes in risk premiums.

An investigation of the relative importance of these channels is complicated by the fact that we donot have high frequency data on firms’ profits and investments and because we do not have a gener-ally accepted measure of equity risk premiums.

In an attempt to nevertheless shed some light on this issue, we study the investment behavior offirms in our sample. If the investment of financially dependent firms responds disproportionately tomonetary shocks, this would suggest that the channel operates through a change in future expectedprofits. If it does not, then this would suggest that the risk premium channel is more relevant. We testthis in a regression model that relates firm investment to the interaction between the US monetarypolicy shock and our measure of financial dependence.

Because we only have annual investment data for firms in our sample, we construct an annual mea-sure of monetary policy shocks by taking the sum of individual monetary shocks over each of theFOMC meetings in a given year. We measure firm investment using the ratio of capital expenditureduring the year to the stock of fixed capital at the beginning of the year. In a second specificationwe modify the definition of firm investment to include assets from acquisitions and abstract fromthe disposal of fixed assets. In both specifications, we include country-year and firm fixed effects,and correct standard errors for clustering at both sector and year levels. The results are presentedin Table 6.

If the investment channel is an important driver of our finding that stock prices of financiallydependent firms respond disproportionately to monetary shocks, then we should find a negativeand significant coefficient on the interaction between monetary shocks and financial dependence inthese investment regressions. Instead, we find that the interaction between the annual monetaryshock variable and financial dependence enters negatively, as expected, but that the coefficient is

Table 6The impact of monetary policy shocks on firm-level investment. Dependent variable in Column 1 is firm investment, i.e., the ratioof capital expenditure during the year to the stock of fixed capital at the beginning of the year. In Column 2, we modify thedefinition of firm investment to include assets from acquisitions and abstract from the disposal of fixed assets. Annual monetaryshock is the sum of monetary policy shocks at each FOMC meeting over a year. Financial dependence is industry’s intrinsicdependence on external finance for investment based on Rajan and Zingales (1998). Standard errors are clustered by sector andyear. Robust standard errors in brackets. ��p < 0.01, ��p < 0.05, �p < 0.1.

(1) (2)

Annual monetary shock � financial dependence �0.010 �0.026[0.030] [0.040]

Firm fixed effects Y YCountry-year fixed effects Y YObservations 105,152 105,152R-squared 0.42 0.38


not significantly different from zero. This suggests that our results are more consistent with the riskpremium channel. These results are consistent with Bernanke and Kuttner (2005) who, using a vectorautoregression model to decompose equity returns find that monetary policy affects stock prices insignificant part by affecting equity risk premiums as opposed to changes in expected real interestrates.

4. Conclusions

This paper studies global stock price responses to US monetary policy shocks using a dataset of20,121 firms across 44 countries over the period 1990 to 2008. We find that stock prices tend to in-crease (decrease) following unexpected monetary loosening (tightening). This impact is more pro-nounced for sectors that depend on external financing, and for countries whose domestic monetarypolicy is more aligned with that of the United States.

The advantage of our asymmetric, cross-sectional identification strategy of estimating differentialeffects of monetary shocks across firms that differ in external financial dependence is that it allows forthe control of unobserved time-invariant effects that simultaneously affect monetary policy as well asa firm’s stock price, thereby alleviating concerns about endogeneity and simultaneity bias. Our inter-national setting also helps to sharpen identification because it is unlikely that US monetary policy re-sponds to (technological) developments in foreign countries.

The economic effects of our results are significant. For example, an unexpected policy rate decreaseof 5 basis points (equal to its interquartile range) is associated with a stock price response that is 6basis points greater for firms whose financial dependence is at the 75th percentile (the Constructionmachinery industry) relative to firms whose financial dependence is at the 25th percentile (the Bev-erages industry). This is a significant effect compared to the average stock market return around FOMCdates of 19 basis points.

The evidence in this paper contribute to the debate about the link between monetary policy andasset prices by showing that prices in stock markets, one of the key financial markets, respondstrongly to monetary shocks. These findings provide new evidence that financial frictions play animportant role in the transmission of monetary policy to the real economy.

Acknowledgments

We would like to thank Viral Acharya (the Editor), two anonymous referees, Geert Bekaert, OlivierBlanchard, Stijn Claessens, Linda Goldberg, Kenneth Kuttner, Claudio Raddatz, David Romer, and sem-inar participants at the 2011 AEA meetings and the International Monetary Fund for comments or sug-gestions, and Jeanne Verrier and Mohsan Bilal for excellent research assistance. The findings,interpretations, and conclusions expressed in this paper are entirely those of the authors. They shouldnot be attributed to the International Monetary Fund.


Appendix A. List of countries

This table lists the total number of firms in each of the 44 countries in our sample over the period1990–2008.

Country
Number of firms
Argentina
51 Australia 1701 Austria 108 Belgium 116 Brazil 236 Canada 1662 Chile 79 China 1387 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 3061 Korea, Republic of 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 2216 Total 20121

Appendix B. FOMC meetings, change in federal funds rate, and monetary shock (in basis points)

This table lists the dates of the FOMC meetings over the period July 1990 to December 2008, the actual change in the federal funds rate on those dates,and the unexpected monetary shocks on those dates.

Date Actual Shock Date Actual Shock Date Actual Shock Date Actual Shock

13-July-90 �25 �14 26-March-96 0 �3 22-August-00 0 �2 10-November-04 25 029-October-90 �25 �31 21-May-96 0 0 3-October-00 0 0 14-December-04 25 014-November-90 �25 4 3-July-96 0 �5 15-November-00 0 0 2-February-05 25 07-December-90 �25 �27 20-August-96 0 �4 19-December-00 0 5 22-March-05 25 018-December-90 �25 �21 24-September-96 0 �13 3-January-01 �50 �38 3-May-05 25 08-January-91 �25 �18 13-November-96 0 0 31-January-01 �50 0 30-June-05 25 01-February-91 �50 �25 17-December-96 0 1 20-March-01 �50 6 9-August-05 25 08-March-91 �25 �16 5-February-97 0 �3 18-April-01 �50 �43 20-September-05 25 130-April-91 �25 �17 25-March-97 25 3 15-May-01 �50 �8 1-November-05 25 246-August-91 �25 �15 20-May-97 0 �11 27-June-01 �25 8 13-December-05 25 013-September-91 �25 �5 2-July-97 0 �2 21-August-01 �25 2 31-January-06 25 031-October-91 �25 �5 19-August-97 0 �1 2-October-01 �50 �7 28-March-06 25 06-November-91 �25 �12 30-September-97 0 0 6-November-01 �50 �10 10-May-06 25 �16-Dec-91 �25 �9 12-November-97 0 �4 11-December-01 �25 0 29-June-06 25 �220-December-91 �50 �28 16-December-97 0 �1 30-January-02 0 2 8-August-06 0 �49-April-92 �25 �24 4-February-98 0 0 19-March-02 0 �3 20-September-06 0 02-July-92 �50 �36 31-March-98 0 0 7-May-02 0 0 25-October-06 0 04-September-92 �25 �22 19-May-98 0 �3 26-June-02 0 �2 12-December-06 0 04-February-94 25 12 1-July-98 0 �1 13-August-02 0 3 31-January-07 0 022-March-94 25 �3 18-August-98 0 1 24-September-02 0 2 21-March-07 0 018-April-94 25 10 29-September-98 �25 0 6-November-02 �50 �19 9-May-07 0 017-May-94 50 13 15-October-98 �25 �26 10-December-02 0 0 28-June-07 0 06-July-94 0 �5 17-November-98 �25 �6 29-January-03 0 1 7-August-07 0 316-August-94 50 14 22-December-98 0 �2 18-March-03 0 5 18-September-07 �50 �1527-September-94 0 �8 3-February-99 0 0 6-May-03 0 4 31-October-07 �25 �215-Nov-94 75 14 30-March-99 0 0 25-June-03 �25 15 11-December-07 �25 120-December-94 0 0 18-May-99 0 �4 12-August-03 0 0 30-January-08 �50 �101-February-95 50 5 30-June-99 25 �4 16-September-03 0 0 18-March-08 �75 1728-March-95 0 10 24-August-99 25 2 28-October-03 0 0 30-April-08 �25 �523-May-95 0 0 5-October-99 0 �4 9-December-03 0 0 25-June-08 0 �36-July-95 �25 �1 16-November-99 25 9 28-January-04 0 0 5-August-08 0 �122-August-95 0 0 21-December-99 0 2 16-March-04 0 0 16-September-08 0 626-September-95 0 0 2-February-00 25 �5 4-May-04 0 �1 29-October-08 �50 �4315-November-95 0 6 21-March-00 25 �3 30-June-04 25 �1 16-December-08 �100 �1219-December-95 �25 �10 16-May-00 50 5 10-August-04 25 231-January-96 �25 �7 28-June-00 0 �2 21-September-04 25 2

L.Laeven,H.Tong

/J.Finan.Intermediation

21(2012)

530–547

545


Appendix C. Correlation between US and local money market rates

This table reports the correlation of monthly money market rates between the US and the individualcountries over the period 1990 to 2008 using data on money market rates from the InternationalFinancial Statistics database of the International Monetary Fund.

Country
Correlation
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 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


References

Ammer, J., Vega, C., Wongswan, J., 2010. International transmission of US monetary policy shocks: evidence from stock prices. J.Money, Credit, Banking 42, 179–198.

Ashcraft, A., Campello, M., 2007. Firm balance sheets and monetary policy transmission. J. Monet. Econ. 54, 1515–1528.Baker, M., Stein, J., Wurgler, J., 2003. When does the market matter? Stock prices and the investment of equity-dependent firms.

Quart. J. Econ. 118, 969–1005.Bekaert, G., Hoerova, M., Lo Duca, M., 2010. Risk, Uncertainty and Monetary Policy. NBER Working Paper No. 16397.Bernanke, B., 2010. Monetary Policy and the Housing Bubble. Speech Delivered at the Annual Meeting of the American Economic

Association, Atlanta, Georgia, January 3, 2010.Bernanke, B., Blinder, A., 1992. The federal funds rate and the channels of monetary transmission. Amer. Econ. Rev. 82, 901–921.Bernanke, B., Gertler, M., 1995. Inside the black box: the credit channel of monetary policy transmission. J. Econ. Perspect. 9, 27–

48.Bernanke, B., Gertler, M., 2001. Should central banks respond to movements in asset prices? Amer. Econ. Rev. 91, 253–257.Bernanke, B., Kuttner, K., 2005. What explains the stock market’s reaction to Federal Reserve policy? J. Finance 60, 1221–1257.Bernanke, B., Mihov, I., 1998. Measuring monetary policy. Quart. J. Econ. 113, 869–902.Boudoukh, J., Richardson, M., Whitelaw, R., 1994. Industry returns and the Fisher effect. J. Finance 49, 1595–1616.Braun, M., 2003. Financial Contractibility and Asset Hardness. Mimeo, Department of Economics, Harvard University.Braun, M., Larrain, B., 2005. Finance and the business cycle: international, inter-industry evidence. J. Finance 60, 1097–1128.Campello, M., 2002. Internal capital markets in financial conglomerates: evidence from small bank responses to monetary

policy. J. Finance 57, 2773–2805.Campello, M., Chen, L., 2010. Are financial constraints priced? Evidence from firm fundamentals and stock returns. J. Money,

Credit, Banking 42, 1185–1198.Cetorelli, N., Goldberg, L., forthcoming. Banking globalization and monetary transmission. J. Finance.Claessens, S., Laeven, L., 2003. Financial development, property rights, and growth. J. Finance 58, 2401–2436.Ehrmann, M., Fratzscher, M., 2004. Taking stock: monetary policy transmission to equity markets. J. Money, Credit, Banking 36,

719–737.Ehrmann, M., Fratzscher, M., 2009. Global financial transmission of monetary policy shocks. Oxford Bull. Econ. Statist. 71, 739–

759.Fama, E., 1981. Stock returns, real activity, inflation, and money. Amer. Econ. Rev. 71, 545–565.Fama, E., Schwert, W., 1977. Asset returns and inflation. J. Finan. Econ. 5, 115–146.Fisman, R., Love, I., 2007. Financial dependence and growth revisited. J. Europ. Econ. Assoc. 5, 470–479.Gertler, M., Gilchrist, S., 1994. Monetary policy, business cycles, and the behavior of small manufacturing firms. Quart. J. Econ.

109, 309–340.Gomes, J., Yaron, A., Zhang, L., 2006. Asset pricing implications of firms’ financing constraints. Rev. Finan. Stud. 19, 1321–1356.Griffin, J., 2002. Are the Fama and French factors global or country specific? Rev. Finan. Stud. 15, 783–803.Gürkaynak, R., Sack, B., Swanson, E., 2005. Do actions speak louder than words? The response of asset prices to monetary policy

actions and statements. Int. J. Central Banking 1, 55–93.Hausman, J., Wongswan, J., 2011. Global asset prices and FOMC announcements. J. Int. Money Finance 30, 547–571.Henry, P., 2002. Is disinflation good for the stock market? J. Finance 57, 1617–1648.Jiménez, G., Ongena, S., Peydró, J., Saurina, J., 2009. Hazardous Times for Monetary Policy: What do Twenty-three million Bank

Loans Say about the Effects of Monetary Policy on Credit Risk? Mimeo, Tilburg University.Kashyap, A., Stein, J., 1994. Monetary policy and bank lending. In: Mankiw, G. (Ed.), Monetary Policy. University of Chicago Press,

Chicago, IL, pp. 221–256.Kashyap, A., Stein, J., 2000. What do a million observations on banks say about the transmission of monetary policy? Amer. Econ.

Rev. 90, 407–428.Kashyap, A., Stein, J., Wilcox, D., 1993. Monetary policy and credit conditions: evidence from the composition of external

finance. Amer. Econ. Rev. 83, 78–98.Kashyap, A., Lamont, O., Stein, J., 1994. Credit conditions and the cyclical behavior of inventories. Quart. J. Econ. 109, 565–592.King, R., Levine, R., 1993. Finance and growth: Schumpeter might be right. Quart. J. Econ. 153, 717–738.Kroszner, R., Laeven, L., Klingebiel, D., 2007. Banking crises, financial dependence, and growth. J. Finan. Econ. 84, 187–228.Kuttner, K., 2001. Monetary policy surprises and interest rates: evidence from the federal funds futures market. J. Monet. Econ.

47, 523–544.Lamont, O., Polk, C., Saá-Requejo, J., 2001. Financial constraints and stock returns. Rev. Finan. Stud. 14, 529–554.Livdan, D., Sapriza, H., Zhang, L., 2009. Financially constrained stock returns. J. Finance 64, 1827–1862.Mackinlay, C., 1997. Event studies in economics and finance. J. Econ. Lit. 35, 13–39.Mishkin, F., 2009. Is monetary policy effective during financial crises? Amer. Econ. Rev. 99, 573–577.Peek, J., Rosengren, E., 2000. Collateral damage: effects of the Japanese bank crisis on real activity in the United States. Amer.

Econ. Rev. 90, 30–45.Piazzesi, M., Swanson, E., 2008. Futures prices as risk-adjusted forecasts of monetary policy. J. Monet. Econ. 55, 677–691.Raddatz, C., 2006. Liquidity needs and vulnerability to financial underdevelopment. J. Finan. Econ. 80, 677–722.Rajan, R., Zingales, L., 1998. Financial dependence and growth. Amer. Econ. Rev. 88, 559–586.Reinhart, C., Rogoff, K., 2004. The modern history of exchange rate arrangements: a reinterpretation. Quart. J. Econ. 119, 1–48.Romer, C., Romer, D., 2004. A new measure of monetary shocks: derivation and implications. Amer. Econ. Rev. 94, 1055–1084.Taylor, J., 2007. Housing and Monetary Policy. Paper Presented at the 2007 Jackson Hole Conference, August.Thorbecke, W., 1997. On stock market returns and monetary policy. J. Finance 52, 635–654.Wongswan, J., 2006. Transmission of information across international equity markets. Rev. Finan. Stud. 19, 1157–1189.

us monetary shocks and global stock prices

Documents