An examination of Eurozone stock markets’ response to Fed and ECB monetary policy announcements

An examination of Eurozone stock markets’

response to Fed and ECB monetary policy

announcements

by

Peter Spijkman

University of Groningen

Faculty of Economics & Business

A

This research examines the relationship between Eurozone stock market indices and mone-tary policy announcements by the U.S. Federal Reserve and the European Central Bank over the period January 1999 through December 2007, utilizing the event-study methodol-ogy. In addition, cross-country analysis is performed to detect and explain asymmetries in the response of Eurozone stock markets to monetary policy surprises. The research paper finds a significant negative relationship between Eurozone stock market indices and U.S. monetary policy surprises, while there is no significant relationship with domestic mone-tary policy surprises by the ECB. Furthermore, cross-country analysis shows that the mag-nitude of the response of Eurozone stock markets to Fed and ECB monetary policy is posi-tively related with the size and liquidity of the financial market. Countries with a relaposi-tively high degree of financial integration with the U.S. show a significantly stronger response to U.S. monetary policy surprises, while a higher degree of real integration with the U.S. is associated with a weaker response to U.S. monetary policy surprises.

Keywords: Monetary policy, Eurozone stock market indices, event study

T

C

2.2.1 Model and Data 10

2.2.2 Methodology 13

3 Methodology and Data 19

3.1 Baseline Model 21

3.2 Measure of Monetary Policy Surprise 22

3.3 Cross-Country Analysis 24

3.4 Robustness 26

4 Results 28

4.1 Baseline Results 28

4.2 Cross-Country Results 33

4.3 Robustness of the Results 37

5 Conclusion 39

References 42

L

A

Fed : Federal Reserve

ECB : European Central Bank

LIBOR : London Inter-Bank Offered Rate FOMC : Federal Open Market Committee CRSP : Center for Research in Security Prices

ATX : Austrian Traded Index

BEL 20 : stock index (Belgium) OMX Helsinki 25 : stock index (Finland)

CAC 40 : Compagnie des Agents de Change-40 (French stock index) DAX 30 : Deutsche Aktienindex (German stock index)

FTSE/ATHEX 20 : stock index (Greece)

ISEQ 20 : Irish Stock Exchange equity index MIB 30 : Milano Indice Borsa (Italian stock index) SE General : stock index (Luxembourg)

AEX : Amsterdam Exchange Index (Dutch stock index) PSI-20 : Portuguese Stock Index

IBEX 35 : stock index (Spain)

CET : Central European Time

GDP : Gross Domestic Product

IMF : International Monetary Fund

1

I

The impact of monetary policy on stock markets is a much investigated topic in financial and monetary economics. However, most research is focused on the U.S. monetary envi-ronment and its effect on U.S. assets. This Master of Science thesis focuses on the response of Eurozone stock markets on monetary policy announcements from the U.S. Federal Re-serve (Fed) and the European Central Bank (ECB) over the period from 1999 through 2007, in particular unexpected changes in its main instrument. The Fed uses the federal funds target rate as its main monetary policy instrument, while Eurozone’s ECB’s mone-tary policy instrument is the main refinancing operations rate. Because this research paper examines the impact of monetary policy announcements on stock markets, the event-study methodology is applied. Together with daily frequency data it is possible to analyze the impact of monetary policy surprises on stock markets. Additionally, a cross-country analy-sis is performed to explore possible asymmetries in the response of Eurozone stock market to the monetary policy surprises.

tensive research to date has been performed by Hausman and Wongswan (2006), who ex-amined the effect of U.S. monetary policy announcement surprises on stock market indi-ces, interest, and exchange rates in 49 countries. However, they have not examined the impact of domestic monetary policy surprises.

An increasing number of research papers in this research field differentiate between antici-pated and unanticiantici-pated news announcements (e.g., Bomfim & Reinhart, 2000; Bomfim, 2003; Ehrmann & Fratzscher, 2004). The line of reasoning behind this is that the market’s expectations are already discounted in the market price. Hence, a news announcement which is in line with the expectations of the announcement will not be interesting for the market, and thus will not have an effect on market prices. The larger the difference be-tween the expectation and the actual announcement, the larger its effect on market returns. Gürkaynak, Sack, and Swanson (2004) provide evidence that the surprise element consists of two factors: the target surprise, and the path surprise. The target surprise captures the current target rate, while the path surprise is related to future monetary policy. They find that U.S. equity indexes react only to the target surprise. Consequently, this research will not incorporate the statements accompanying the monetary policy announcement, and only considers the target rate.

there is not a comparable futures rate that tracks the ECB’s main refinancing rate. There-fore, this research paper uses London Inter-Bank Offered Rates (LIBOR) to measure the expectations of the monetary policy announcement, and to calculate the surprise element in the monetary policy announcement. LIBOR is a reference rate on which banks offer to lend unsecured funds to other banks in the interbank market. It is used by market partici-pants to fix the costs of borrowing in financial markets around the world, and is fixed daily for the currencies, amongst others, euro and U.S. dollar with maturity ranging from over-night up to one year. Since this research examines both U.S. as well as Eurozone monetary policy, LIBOR is used rather than Euribor which is only quoted in euro’s. In the compari-son by Gürkaynak et al. (2002), Eurodollar deposit rates gave the second-best results as measure for monetary policy expectations, after federal funds futures. However, a few other market based measures are of comparable quality, probably due to the high degree of integration of these markets. Nevertheless, the accessibility and availability of LIBOR data are the reasons to opt for LIBOR instead of the other measures in this research paper. Furthermore, this research paper performs a cross-country analysis to explore asymmetries in the response of Eurozone stock market indices to Fed and ECB monetary policy sur-prises. Previous research has shown that asymmetries in stock market response to mone-tary policy is related to the degree of financial integration, real integration, and financial market development (Wongswan, 2006; Hausman & Wongswan, 2006; Ehrmann & Fratzscher, 2006). Therefore, this research paper investigates whether financial integration, real integration, and financial market development are associated with the response of Eu-rozone stock markets to Fed and ECB monetary policy.

The objective of this research is to examine the relationship between Eurozone stock mar-kets and monetary policy announcements by the Fed and the ECB. Moreover, potential asymmetries in the response of Eurozone stock markets to monetary policy surprises by the Fed or the ECB are explored. The conceptual model in Appendix B presents a graphical explanation of this research. The following research questions have been formulated in order to guide this research, and are divided into one main research question, and two sub questions.

The main research question is:

The sub research-questions are:

- Do Eurozone stock market indices respond asymmetrically to monetary policy sur-prises by the Fed and the ECB?

a) If so, to which central bank’s monetary policy surprises are Eurozone stock market indices’ response stronger?

- Do financial market development, financial integration, or real integration explain asymmetries in the response of Eurozone stock market indices to monetary policy sur-prises?

2

T

B

This section describes the existing literature on the subject of the effect of monetary policy on stock markets. It has been divided into two subsections. First, the effect of monetary policy on equity markets from a theoretical point of view is explained. Thereafter, empiri-cal studies which have already investigated the relationship are described in terms of their data, methodology, and findings.

2.1 Theoretical Relationship

The theoretical relationship refers to the macroeconomic relationship between monetary policy and equity markets, more specifically why and how equity markets are affected by monetary policy. Through monetary policy central banks have the ability to influence the money supply and interest rates to achieve economic goals. The Fed and the ECB have three tools through which they can conduct monetary policy of which the open market op-erations is the most important one. The other two instruments are the discount rate, and the reserve requirements. The open market operations entails the purchases and sales of U.S. treasury and federal agency securities, and its goal is to target a specific interest rate; the Fed targets the federal funds rate, and the ECB targets the main refinancing rate. Adjust-ments in these particular rates affect other long- and short-term interest rates, exchange rates, and eventually macroeconomic variables such as inflation, economic growth, and unemployment. On the other hand, equity prices are claims on future economic output, and thus changes in monetary policy can affect stock prices. More specifically, as the value of a stock represents the sum of discounted future dividends, monetary policy influences stock prices through the discount rate and expected future earnings (Sellin, 2001).

future dividends, while “monetary policy has little to with future expected real interest rates” (p. 1968). Bernanke and Kuttner (2005) confirm this finding in that they also ex-clude future expected real interest rates as a source for the stock market’s response to an unanticipated change in monetary policy, which leaves two other possible sources for the market reaction: expected future dividends, or expected excess equity returns. However, they are not conclusive about which of these two is the primary source.

2.2 Empirical Studies

This subsection deals with the extensive amount of empirical research papers which have contributed to the field of research by investigating the effect of stock markets to monetary policy. These research papers differ from each other in terms of model and data, methodol-ogy, and findings. A list of research papers which have investigated the relationship be-tween stock markets and monetary policy previously is summarized in Table I at the end of this section.

2.2.1 Model and Data

The basic research model used in the field of research illustrates a relationship between monetary policy and some type of stock index. Waud’s (1970) paper was the first attempt to measure the effect of stock markets to monetary policy. However, since Waud’s (1970) paper the base research model has been adjusted to extend knowledge in the relationship between monetary policy and stock markets.

How-but reported similar results. Research papers using the federal funds rate as monetary pol-icy instrument most often use the federal funds futures rate to measure the surprise element in the monetary policy announcement (e.g., Bomfim & Reinhart, 2000; Bomfim, 2003; Davig & Gerlach, 2006). Two other research papers examine the effect of stock markets to monetary policy in Japan (Honda & Kuroki, 2006) and the U.K. (Bredin et al., 2007). Con-sequently, these paper use the 3-month EuroYen futures rate and the 3-month Sterling fu-tures rate to isolate the surprise element in the monetary policy announcement.

Smirlock and Yawitz (1985) took an alternative approach and made a distinction between “technical” and “non-technical” discount rate changes. Technical changes are endogenous and should not affect market prices in an efficient market, where non-technical changes are unexpected. They find that the market reacts more strongly to a nontechnical rate change. Likewise, Gürkaynak et al. (2004) made a distinction between the “target surprise” and “path surprise” using information from the FOMC statements along the federal funds rate decision. In their case, the target surprise refers to the surprise element related to the cur-rent change in the federal funds rate, while the path surprise refers to possible future changes in the federal funds rate up to a horizon of one year. They found that stock prices, in their paper the S&P 500, only react significantly to the target factor, and not to the path factor.

So, it is important to isolate the surprise element in the monetary policy announcement to capture the true effect of monetary policy on stock market prices. Therefore, this research paper uses LIBOR because of its quotation in both euro and U.S. dollar, and its availability and accessibility of the data, while the quality of the results is not weakened (Gürkaynak et al., 2002).

In addition to the identification of the surprise element of the monetary policy announce-ment, asymmetries in the response of stock prices to monetary policy have been investi-gated. These asymmetries can be found on firm, sector, and country level.

announcements than other industries. On the other hand, less cyclical industries are less sensitive to the announcements. Moreover, sectors with typical capital intensive forms of investments, like the oil industry, are extremely sensitive to monetary policy shocks (Bredin et al., 2007). In addition, Bernanke and Kuttner (2005) found the high-tech and telecommunications sectors as the most responsive sectors, later confirmed by Ehrmann and Fratzscher (2006) who found the information technology sector as most responsive. Both research papers found the utility sector least responsive, and showed a relationship between the industry beta’s and the response to monetary policy surprises. In general, in-dustries with a high market beta are more responsive to monetary policy surprises.

Asymmetry in response to monetary policy can also be due to firm characteristics. The main characteristic used in these studies is the firm’s size in terms of market capitalization. Thorbecke (1997) divided the broad CRSP index into ten size portfolios and finds that monetary policy shocks have a larger effect on small firms than on large firms. Thorbecke and Coppock (1996) confirm this finding, however they find that small firms are not helped in periods of monetary expansion, opposite to large firms which do benefit. A somewhat similar finding was done by Madura and Schnusenberg (2000), who performed an analysis of monetary policy changes on the stock prices of commercial banks only, and found large banks to benefit to a greater extend to interest rate decreases than relatively small banks. In addition, the same paper also investigated the role of capital ratio. They found high capital ratio banks to be less sensitive to monetary policy changes, which is consistent with the hypothesis that banks with a high capital ratio have a lower degree of leverage, so the bank’s value should be less affected to changes in net income.

exchange rate regime is more responsive to U.S. monetary policy. Moreover, Ehrmann & Fratzscher (2006) also find countries with open and relatively liquid financial markets to respond more strongly to U.S monetary policy shocks than less integrated countries. How-ever, in contrast to Wongswan’s (2006) and Hausman and Wongswan’s (2006) findings it is the degree of integration with the entire world, and not the bilateral integration with the U.S., that determines the magnitude of the stock market response to U.S. monetary policy. Conclusively, asymmetries in the response of stock market to monetary policy surprises can be attributed to the extend of integration with other countries, and the openness and liquidity of the financial markets. In contrast to previous research papers which investi-gated the response of equity prices to only one monetary policy instrument, this research paper examines the response of Eurozone stock markets to both Fed and ECB monetary policy surprises. This approach enables a comparison of Eurozone stock markets’ response to domestic and foreign monetary policy surprises. In addition, this research explores asymmetries in the response of stock markets and its relationship with financial develop-ment, financial integration, and real integration. Hereby, two new proxies for financial de-velopment are used which measure liquidity in the financial market.

2.2.2 Methodology

24-hour window around the monetary policy announcement, however some recent studies used intraday data instead to narrow the event-window even further (e.g., Wongswan, 2006). Although using intraday data improves the power of the statistical model (R2), the stock market response is similar to that obtained using daily data (Gürkaynak et al., 2004).

Another solution to these two problems is the use of alternative statistical methods. The common methodology applied in the field of research is the event study. This method views the monetary policy shock as an exogenous variable which affects stock market re-turns. Thus, it is assumed that stock market returns respond to the monetary policy shock, and not the other way around. In addition, the event-study methodology assumes the error-term to be orthogonal with monetary policy changes. So, the error error-term (the proportion of variance in stock market returns not explained by monetary policy announcements) is as-sumed to be unrelated with monetary policy changes. This assumption may be violated when other shocks affect stock market prices.

To circumvent this assumption Rigonbon and Sack (2004) use a heteroskedastic-based estimator which exploits the heteroskedasticity common for exogenous monetary policy announcements. Unlike the event-study, this heteroskedastic-based estimator does not as-sume the monetary policy surprise to dominate all other shocks to the stock market, and only requires a shift in the relative importance of the shocks. Craine and Martin (2003) use a multivariate factor model where security prices react to a market wide systematic and security specific idiosyncratic sources of risk, and where monetary policy and other news announcements are the factors. Although these approaches require a weaker set of assump-tions, it generates similar results to those studies using the event-study methodology. Therefore, the event-study methodology is applied in this research paper and will be de-scribed in the methodology section.

Table I: Research Papers Study Monetary Policy in-strument* No. of Observa-tions

Stock market** Data interval Period(s) *** Type of Study**** Expected / Unexpected Findings

Waud (1970) DR 25 S&P500 Daily 1952 – 1967 Event Study None Negative

relationship

Pearce & Roley (1983) M1 204 DJIA Daily 1977:9 – 1979:10 &

1979:10 – 1980:1 & 1980:1 – 1982:1

Event Study Survey data Negative relation-ship only for unanticipated Smirlock & Yawitz

(1985)

DR 36 NYSE VW Daily 1975 – 1979 &

1979 – 1982

Event study Federal Reserve Announcement

No relationship & negative effect for non-technical changes Jensen & Johnson

(1993) DR 75 CRSP EW, financial EW Daily 1962 – 1979:10 & 1979:10 – 1982:10 & 1982:10 – 1990 Event study (CMAR) Federal Reserve Announcement Negative relation-ship for all 3

periods Thorbecke & Alami

(1994)

FFR 76 DJIA, DJCA,

SPCA

Daily 1974:9 – 1979:9 Event study None Negative

relationship Jensen & Johnson

(1995) DR 78 CRSP EW, VW & financial EW Daily 1962 – 1979 & 1979 – 1991 Event study (CMAR) Federal Reserve Announcement Negative relation-ship for both

periods Thorbecke & Coppock

(1996)

FFR, NBR 70 10 VW NYSE

portfolios

Daily 1974:9 – 1979:9 & 1982:8 – 1987:9

NLSUR None Negative

Study Monetary Policy in-strument* No. of Observa-tions

Stock market** Data interval Period(s) *** Type of Study**** Expected / Unexpected Findings

Jensen, Johnson & Bauman (1997)

DR 73 CRSP EW, 16 EW

industry indices

Daily 1968:8 – 1991 Event study None Negative

relation-ship for both short- and

long-term

Patelis (1997) FFR, NBR 393 CRSP VW Monthly 1962:1 – 1994:11 Event study,

VAR

None Negative relation-ship for FFR only

Thorbecke (1997) FFR 116 DJIA, DJCA Daily 1974 – 1994 VAR, Event

study, NLSUR

None Negative

relationship Conover, Jensen &

Johnson (1999)

DR 363 10 European, U.S.,

UK, Japan, Can-ada, NZ, SA

Monthly 1956:1 – 1995:12 OLS

regression

None Negative

relationship

Bomfim & Reinhart (2000)

FFR 98 S&P500, Nasdaq Daily 1989:5 – 1994:2 &

1994:2 – 1998:12

Event study Survey data and future contract

No relationship

Madura & Schnusen-berg (2000) FFR, DR FFR: 148 DR: 60 CRSP commercial banks only Daily 1974:9 – 1979:9 & 1979:10 – 1987:8 & 1987:9 – 1996:12

Event study None Negative

relation-ship for all 3 periods

Bomfim (2003) FFR 2414 S&P500 Daily 1989:6 – 1998:12 Event study Federal funds

futures rate

Negative relationship Craine & Martin

(2003)

FFR 144 CRSP VW Daily 1988:10 – 2001:12 Factor model Federal funds

futures rate

Negative relationship

Bentzen et al (2004) FFR 71 SPY Intraday 1995:1 – 2002:12 Event study Federal funds

futures rate

Study Monetary Policy in-strument* No. of Observa-tions

Stock market** Data interval Period(s) *** Type of Study**** Expected / Unexpected Findings

Bredin et al (2004) FFR 114 S&P500, ISEQ Daily 1989:6 – 2003:6 Event study Federal funds

futures rate Negative relationship Ehrmann and Fratzscher (2004) FFR 78 S&P500 + stocks therein

Daily 1994:2 – 2003:2 Event study Survey data Negative

relationship Rigobon and Sack

(2004) 3-month euro-dollar rate 73 DJIA, S&P500, NASDAQ, Wilshire 500 Daily 1994:1 – 2001:11 Heteroskedas-ticity-based estimator 3-month Euro-dollar rate Negative relationship

Bernanke and Kuttner (2005)

FFR 131 CRSP VW Daily 1989:5 – 2002:12 Event study Federal funds

futures rate

Negative relationship Davig & Gerlach

(2006)

FFR 80 S&P500 Intraday 1994 – 2003:12 Event study Federal funds

futures rate

Negative relation-ship, except for 1998:9 – 2002:9 Ehrmann & Fratzscher

(2006)

FFR 93 50 countries Daily 1994:2 – 2004:12 Event study Federal funds

futures rate 30 minute window Negative relationship Hausman & Wongswan (2006)

FFR 94 49 countries Daily 1994:2 – 2005:3 Event study Federal funds

futures rate

Negative relationship Honda & Kuroki

(2006)

Bank of Japan Call rate

55 Nikkei225, TOPIX Daily 1989:7 – 2001:3 Event study 3-month

Euro-Yen futures Negative relationship Wongswan (2006) FFR 53 S&P500, FTSE100, 3 EU, 9 Asia, 3 LA

Intraday 1998:9 – 2004:11 Event study Federal funds futures rate

Study Monetary Policy in-strument* No. of Observa-tions

Stock market** Data interval Period(s) *** Type of Study**** Expected / Unexpected Findings

Robitaille & Roush (2006)

FFR 43 IBOVESPA Intraday 1999:2 – 2005:4 Event study Survey data Negative

relationship Bredin et al (2007) 2-week repo

rate (BOE base rate)

unknown FTSE, 16 sector portfolios

Daily 1993:1 – 2004:5 Event study 3-month Sterling futures contract

Negative relationship

* DR = Discount rate; M1 = Money aggregate; FFR = Federal funds rate target; NBR = Non-borrowed reserves

** VW = Value Weight; EW = Equal Weight; DJCA = Dow Jones Composite Average; SPCA = Standard & Poor’s Composite Average; SPY = Standard & Poor’s Depository Receipts

*** Date notation in YYYY : MM; Periods separated by “&”

3

M

D

This paper focuses on monetary policy announcements by the ECB and the Fed, and its effect on Eurozone stock market indices in the period from 1999 through 2007. The mone-tary policy instruments included in this research are the Fed’s federal funds target rate, and the ECB’s main refinancing rate. Data about changes in the federal funds rate and the main refinancing rate are gathered from the websites of the central banks,1 and the source of the measure of the expected decision (LIBOR) is the website of the British Bankers’ Associa-tion.2 The Fed’s first event day is the FOMC meeting on February 3, 1999 (no change), and the last is the December 11, 2007 one-quarter percent rate cut. The ECB’s first event day is January 7, 1999, and the last is December 6, 2007 (both unchanged). The September 17, 2001 half percentage point rate cuts by the Fed and ECB have been excluded from the analysis which were part of a joint response by central banks following the terrorist attacks on September 11, 2001. In total, there are 74 observations for the Fed, and 111 observa-tions for the ECB, in which the Fed changed its primary policy tool 38 times (23 rate hikes and 15 rate cuts), and the ECB accounts for 22 changes (15 rate hikes and 7 rate cuts). This amount of observations is comparable to the amount of observations in related studies, and specifically in multi-country studies like, for example Ehrmann and Fratzscher (2006) (see Table I).

Daily opening and closing prices have been gathered from January 1999 through Decem-ber 2007 for the following stock market indexes: ATX (Austria), BEL20 (Belgium), OMX Helsinki 25 (Finland), CAC 40 (France), DAX 30 (Germany), FTSE/ATHEX 20 (Greece), ISEQ 20 (Ireland), MIB 30 (Italy), SE General (Luxembourg), AEX (Netherlands), PSI-20 (Portugal), IBEX 35 (Spain), and the Dow Jones EuroStoxx 50 (Eurozone). The Dow Jones Eurostoxx 50 index is included in this research paper to gauge the response of the 50 largest Eurozone companies to monetary policy announcements by the Fed and ECB. Greece entered the European Economic and Monetary Union in January 2001, therefore it has been omitted from analysis until that date. As Slovenia entered the European Economic and Monetary Union in January 2007 its sample size is too small and therefore completely omitted from the analysis. The source for these data is Datastream. Descriptive statistics of the variables are displayed in Table II.

1 _{www.federalreserve.gov and www.ecb.eu} 2

Table II: Descriptive Statistics

Mean St. Dev. Minimum Maximum N

Monetary Policy:

Fed

Raw change .000 .2413 -.50 .50 74

Monetary policy surprise -.0201 .0878 -.4200 .1000 74

ECB

Raw change .0136 .1508 -.50 .50 111

Monetary policy surprise .0002 .0417 -.2328 .1569 111

Index Returns (%): ATX .2571 1.0346 -4.05 3.70 185 BEL 20 _{.0358 .9964 -4.51 4.36} 185 OMX H25 .2475 1.6904 -7.68 7.42 185 CAC 40 _{.0107 1.2127 -5.43 5.61} 185 DAX 30 -.0396 1.2211 -4.33 5.40 185 FTSE/ATHEX 20 .2808 1.0131 -1.93 3.72 144 ISEQ 20 _{.0449 1.2080 -2.63 7.04} 185 MIB 30 -.0001 1.1452 -3.48 4.22 185 SE General _{.1326 1.1325 -3.75 3.65} 185 AEX .0061 1.1818 -5.14 4.64 185 PSI 20 -.0540 .7635 -3.61 3.23 185 IBEX 35 _{.0415 1.1330 -4.27 5.57} 185 DJ EuroStoxx 50 .0304 1.1669 -4.85 4.03 185 Weighted mean .0208 1.0127 -3.61 4.26 185

Equal weight mean .0788 .7849 -2.20083 2.25833 185

Financial Development:

Stock Market Capitalization

(% of GDP) 326.92 200.3884 50.53 1710.68 2178

Total Value Traded (% of GDP) 25.113 20.2013 .11 93.04 2178

Total Value Traded (% of Stock

Market Capitalization) 8.056 5.1787 .01 27.45 2178

Financial Integration:

U.S. Equity Participation (%) 12.282 6.733 2.34 33.71 500

Eurozone Equity Participation (%) 16.683 7.400 1.14 35.02 730

Real Integration:

Trade with U.S. (% of GDP) 1.702 1.407 .22 7.46 872

3.1 Baseline Model

This research paper utilizes the event-study methodology (see, e.g., Brown & Warner, 1980, 1985; MacKinlay, 1997). More specifically, the approach by Bernanke and Kuttner (2005) is followed where over the event window stock market index returns are regressed on the actual change in the monetary policy instruments from the Fed and ECB,

Rt = α + β ∆ίt+ εt (1)

where Rt is the percentage change in the stock market index, ∆ίt is the change in

percent-age points of the monetary policy instrument, and εt is the error term. Following the event

study approach the sample consists of days on which the Fed and ECB changed its mone-tary policy instrument, and days corresponding with regularly scheduled meetings where the monetary policy instrument was unchanged.

This baseline model is used in many research papers from Table I which use the event study methodology. However, earlier research papers have not reported congruent find-ings. Early research papers which only investigated the raw change in monetary policy instruments, like Waud (1970) and Thorbecke and Alami (1994), found a significant nega-tive relationship with stock market returns. On the other hand, more recent research papers found an insignificant relationship between the raw change in monetary policy instrument and stock market returns (Bernanke & Kuttner, 2005; Honda & Kuroki, 2006). Since this research paper includes recent data, results similar to findings from other recent research papers are expected. Therefore, the corresponding hypothesis becomes:

H1: Eurozone stock market returns are not related to raw changes in Fed and ECB mone-tary policy instruments.

fol-lows. For ECB’s decisions Rt is the difference between the closing price and the opening

price on the day of the announcement (day t), and for Fed’s decisions Rt is the difference

between the opening price on the day after the announcement (day t+1) and the closing price on the day of the announcement (day t).

However, two problems with the opening price data arise. First, opening data is not avail-able for all indices and for every period. In these cases, for ECB announcements the open-ing price is replaced with the closopen-ing price on day t-1, and for Fed announcements the opening price is replaced with the closing price on day t+1. Second, the opening prices of ATX (Austria), OMX Helsinki 25 (Finland), FTSE/ATHEX 20 (Greece), and SE General (Luxembourg) are equal or virtually equal to the closing price of the day before, which makes it hard to detect a significant response to announcements by the Fed. In these cases the opening price has also been replaced by the closing price on day t+1. For a detailed list of data fixes see Appendix C.

3.2 Measure of Monetary Policy Surprise

To measure the monetary policy surprise this research paper uses LIBOR as it is quoted in euro’s and U.S. dollars, thus applicable to both Fed and ECB monetary policy announce-ments. Market expectations on monetary policy decisions are incorporated in market inter-est rates like LIBOR, and are thus appropriate for isolating the surprise element following:

Ut = Lt+1 - Lt (2)

and,

Et = ∆ίt - Ut (3)

where Ut represents the unanticipated change in the monetary policy instrument,

calcu-lated by the one day change in the 3-months ahead LIBOR. The expected part of the policy announcement, Et is then extracted by calculating the difference between the actual change

LIBOR rates are released 11.00 London time (12.00 CET) each day, thus released prior to both Fed and ECB monetary policy announcements on day t. Hence, the unanticipated change in monetary policy is calculated as the difference between the LIBOR the day after the policy announcement (t+1) and the LIBOR on the day of the announcement (t). The figure in Appendix D presents a graphical display of the opening hours of Eurozone stock markets and release times of the monetary policy decisions and LIBOR.

Using the decomposition between the expected and unexpected part above, the following regression equation is used, which is similar to Bernanke and Kuttner’s (2005) approach:

Rt = α + β1 Ut + β2 Et + εt (4)

Theoretically, stock markets only respond to the unexpected component of the monetary policy announcement as that is the part of the information that is actually new for market participants. Earlier research has shown that a surprise rate cut in the monetary policy in-strument is associated with higher stock prices, and a surprise rate hike is related with lower stock market prices (see Table I). The expected component on the other hand, is what market participants have anticipated and is already discounted in the market price. Therefore, the expected component will not affect stock market prices. Consequently, the following two hypotheses are:

H2a: Eurozone stock market returns are negatively related to the unexpected component of Fed and ECB monetary policy announcements.

H2b: Eurozone stock market returns are not related to the expected component of Fed and ECB monetary policy announcements.

3.3 Cross-Country Analysis

As discussed in the theoretical background, asymmetries occur on a country level where some countries react more strongly to monetary policy surprises than other countries. Pre-vious research has shown that this can be explained by the degree of integration with other countries, and the financial market development of the country. This research paper uses several proxies which are categorized in financial market development, financial integra-tion, and real integration. Some of these proxies have been used in previous research and proved their predictive value (Wongswan, 2006; Hausman & Wongswan, 2006; Ehrmann & Fratzscher, 2006), while this research adds two proxies for financial market develop-ment. Financial market development proxies are used to capture the importance and liquid-ity of the stock market in a country. Ehrmann and Fratzscher (2006) find that countries with relatively open and well developed equity markets and financial sectors show a stronger response to U.S. monetary policy surprises. Also, the higher the degree of finan-cial and real integration, the stronger the stock market response to monetary policy.

Three different proxies for financial market development are used. First, the size of the domestic equity market relative to the country’s GDP, calculated by the ratio of equity market capitalization to GDP. Second, the stock market’s total value traded relative to the GDP. Third, the stock market turnover, or the stock market’s total value traded relative to the size of the equity market. This is calculated by the percentage of the total value traded on the stock market to the stock market capitalization. Equity market capitalization data and total value traded data at monthly frequency are gathered from the World Federation of Exchanges,3 and GDP data is from Datastream. The first proxy, the size of the domestic equity market relative to the country’s GDP, has also been used in previous research. Hausman and Wongswan (2006) found this proxy for the importance of financial sector development to be an important factor in explaining the asymmetries in the response of global stock markets to U.S. monetary policy announcements. Ehrmann and Fratzscher (2006) presented similar results as they found countries with higher stock market capitali-zation to show a stronger response to U.S. monetary policy. The last two proxies, however, have not been used in previous cross-country analysis, and examine whether liquidity of

the financial market is related to the response of stock markets to monetary policy an-nouncements.

For both financial and real integration this research uses only one proxy. For financial in-tegration a proxy for equity participation by foreign investors is used, calculated as the percentage of domestic equity market capitalization owned by foreign investors. Equity market capitalization owned by foreign investors data are from IMF’s Coordinated

Portfo-lio Investment Survey, and available at annual frequency from 2001 through 2006.

How-ever, two countries (Ireland and Luxembourg) show abnormal high values for stock market capitalization owned by foreign investors. For instance, over the complete sample the amount of stock market capitalization owned by Eurozone investors in Luxembourg is at least seven times the total stock market capitalization. Therefore, Ireland and Luxembourg are removed from the sample, and the financial integration proxy is run with data from ten countries. Equity participation by foreign investors has proved to be the most successful estimator for financial integration in explaining cross-country asymmetries (Hausman & Wongswan, 2006). The common proxy for real integration is the ratio of trade (exports and imports) to GDP. Exports and imports data are gathered from IMF’s Direction of Trade

Statistics through Datastream at monthly frequency for the complete sample, and GDP data

is from Datastream. The ratio of trade to GDP is a commonly used proxy in related studies (Wongswan, 2006; Hausman & Wongswan, 2006; Ehrmann & Fratzscher, 2006). These studies find cross-country asymmetries in the response to U.S. monetary policy announce-ments to be significantly related to the ratio of trade to GDP. Although, real integration is found to be less important in explaining cross-country asymmetries than proxies for finan-cial integration (Wongswan, 2006; Hausman & Wongswan, 2006). See Table II for the descriptive statistics of these variables.

To estimate the influence of these factors on asymmetries in the response of stock markets the following regression is used:

Rt = α + β1 Ut + β2 (Ut * Xt)+ εt (5)

where Xt is one of the five proxies used to explain cross-country variance in the response to

As a result of the findings in research papers by Hausman and Wongswan (2006) and Ehrmann and Fratzscher (2006) the following hypotheses are:

H3a: The magnitude of the response of Eurozone stock markets to monetary policy sur-prises is positively related with the degree of financial development.

H3b: The magnitude of the response of Eurozone stock markets to monetary policy sur-prises is positively related with the degree of financial integration.

H3c: The magnitude of the response of Eurozone stock markets to monetary policy sur-prises is positively related with the degree of real integration.

3.4 Robustness

The results from the regression estimations above can be driven by potential problems with the models and data used. Therefore, this research paper conducts several robustness checks on the results to exclude influence of these imperfections on the results in this re-search paper. The first couple of tests for robustness are related to the quality of the data. First, the response of stock market to monetary policy surprises may be related to certain characteristics of the monetary policy announcements. So, this research checks whether stock market response is influenced by, 1) whether there was a change in the monetary policy instrument or not, 2) whether the change at a particular meeting was a directional change or not, and 3) whether the monetary policy surprise was positive or negative. These robustness checks are performed following an approach similar to Ehrmann & Fratzscher (2006):

Rt = α + β1 Ut Dt + β2 Ut (1-Dt) + εt (6)

where Dt is a dummy variable and equals “1” when, 1) there was a change, 2) the change

was a directional change, and 3) the monetary policy surprise was positive. The dummy variable Dt equals “0” otherwise. Then, a t-test is performed to determine whether β1 and

β2 are significantly different from each other.

these dates show differences in stock market returns equation 6 is used, where dummy variable Dt equals “1” when the monetary policy decision date corresponds with a Bank of

England decision, and equals “0” otherwise.

4

R

This section of the research paper presents the results following the methodology presented above. The same order is applied here; baseline results and results with isolation of the surprise element, cross-country asymmetry results, and finally the robustness checks.

4.1 Baseline Results

The results from the baseline model (1) are presented in table III. The table is divided into three samples, so that asymmetries in the response of stock markets are visible: a sample with only Fed monetary policy decisions, a sample with only ECB monetary policy deci-sions, and a sample with monetary policy decisions from both central banks. Moreover, the response of each single stock market index, and a weighted- and equal weight mean stock market return for the three samples are presented.

Table III: Baseline Results

This table shows estimates from the regression of Euzone stock market returns on raw changes in monetary policy actions, following:

Rt = α + β ∆ίt+ εt (1)

The sample consists of monetary policy data from January 1999 through December 2007, excluding the case of September 11, 2001. The sample of the Greek stock market index FTSE/ATHEX 20 starts from January 2001, hence contains 58 Fed cases and 85 ECB cases. Weighted mean is calculated on the basis of stock market capitalization. Significant coefficients at the 5% level are in bold.

FED (N=74) ECB (N=111) Full Sample (N=185)

β sign. R2 β sign. R2 β sign. R2

ATX -.115 .328 .013 -.051 .598 .003 -.085 .254 .007 BEL 20 -.006 .956 .000 .164 .087 .027 .086 .246 .007 OMX H25 -.022 .852 .000 .007 .942 .000 -.007 .926 .000 CAC 40 -.022 .852 .000 .049 .608 .002 .014 .855 .000 DAX 30 .016 .892 .000 .078 .420 .006 .047 .529 .002 FTSE/ATHEX 20 .314* .016 .099 .099 .370 .010 .210* .012 .044 ISEQ 20 .363** .001 .132 .140 .146 .019 .246** .001 .061 MIB 30 -.092 .435 .008 .156 .103 .024 .036 .627 .001 SE General .189 .107 .036 -.017 .863 .000 .077 .298 .006 AEX .094 .426 .009 .091 .344 .008 .081 .277 .006 PSI 20 -.013 .909 .000 -.040 .681 .002 -.030 .687 .001 IBEX 35 -.154 .191 .024 .041 .672 .002 -.050 .501 .002 DJ EuroStoxx 50 .007 .956 .000 .070 .467 .005 .043 .561 .002 Weighted mean -.018 .876 .000 .087 .365 .008 .038 .607 .001

Equal weight mean .071 .546 .005 .091 .334 .008 .073 .323 .005

Table IV shows the response of Eurozone stock markets to the change in the target rate when it is divided into its expected and surprise components, again for both the Fed and ECB separately and for the full sample. The significance is below the coefficient in paren-theses. The table presents two important findings. First, Eurozone stock market indices only respond significantly to monetary policy by the Fed, and insignificantly to monetary policy by the ECB. With the exception of the ATX and the FTSE/ATHEX 20, all Euro-zone stock market indices show a significant response to either the surprise component or the expected component of Fed’s monetary policy announcements. The coefficients of the surprise components are all negative, and the coefficients of the expected components are all positive. A hypothetical one percent surprise cut in the federal funds rate is related with a average 3.39% gain in Eurozone stock market indices. However, Eurozone stock market indices show some asymmetries in the response to Fed monetary policy announcements. For instance, the largest significant movement will be in the IBEX 35 which will gain 6.26%, while the SE General will gain 3.38%. For ECB monetary policy announcements, only the BEL 20 and the MIB 30 show a significant relationship with the expected part of the monetary policy announcement only. Moreover, Eurzone stock market returns are also more explained by U.S. monetary policy than by ECB’s monetary policy. Monetary policy announcements by the Fed explain 24.4% of the variance of one-day Eurozone stock mar-ket indices’ returns, while only 15% of the variance in Eurozone equity prices are associ-ated with ECB’s monetary policy announcements. So, in general Eurozone stock market indices are more strongly related to U.S. monetary policy than to its domestic monetary policy.

Table IV: Surprise Measurement Results

This table shows estimates from the regression of Eurozone stock market returns on the surprise components and expected components of monetary policy actions following:

Rt = α + β1 Ut + β2 Et + εt (4)

The sample consists of monetary policy data from January 1999 through December 2007, excluding the case of September 11, 2001. The sample of the Greek stock market index FTSE/ATHEX 20 starts from January 2001, hence contains 58 Fed cases and 85 ECB cases. Weighted mean is calculated on the basis of stock market capitalization. Significant coefficients at the 5% level are in bold. Significance is below the coefficients in parentheses.

FED (N=74) ECB (N=111) Full Sample (N=185)

FED (N=74) ECB (N=111) Full Sample (N=185) β1 β2 R2 β1 β2 R2 β1 β2 R2 PSI 20 -.507 ** (.000) .308** (.005) .266 -.094 (.333) -.002 (.981) .009 -.277** (.000) .124 (.093) .077 IBEX 35 -.573 ** (.000) .200 (.057) .305 -.062 (.528) .068 (.486) .007 -.304** (.000) .117 (.107) .091 DJ EuroStoxx 50 -.426 ** (.000) .278* (.015) .194 -.059 (.542) .097 (.318) .011 -.076 (.316) .088 (.245) .011 Weighted mean -.506 ** (.000) .302** (.006) .263 -.066 (.493) .118 (.222) .016 -.228** (.002) .168* (.023) .064

Equal weight mean -.451

** (.000) .362** (.001) .244 -.056 (.563) .118 (.225) .015 -.239** (.001) .211** (.004) .080

equally weighted international stock markets sample. Finally, the few papers which exam-ine the relationship outside the U.S. find contradicting results. Honda and Kuroki (2006) found the Japanese stock market to respond insignificantly to monetary policy by the Bank of Japan. Bredin et al. (2007) did find a significant negative response of the FTSE to U.K. monetary policy surprises, however the response was found to be weaker than in U.S. stud-ies. Therefore, a weak or insignificant relationsip of non-U.S. stock markets with domestic monetary policy seems a valid finding, and it can be concluded that U.S. monetary policy plays a dominant role in international stock market returns, more than domestic monetary policy.

Second, Eurozone stock markets show a significant response to both the surprise compo-nents and the expected component of monetary policy announcements by the Fed. The ISEQ 20 even only shows a significant relationship with the expected component, and in-significant with the surprise component. So, hypothesis H2a is supported and hypothesis H2b is rejected. Theory suggest that asset prices do not react to expected part of news as it is already discounted in the market price. However, Bernanke and Kuttner (2005) also re-ported a significant response of the CRSP value weight index to the anticipated part of U.S. monetary policy announcements in addition to the surprise component. Also, the R2 of their model is comparable to the findings in this research paper.

4.2 Cross-Country Results

The results from the cross-country analysis (5) are displayed in table V. This analysis tries to identify possible factors that may influence the response of Eurozone stock market indi-ces to monetary policy surprises by the Fed and the ECB by regressing the stock market index return on one of the proxies for financial development, financial integration, or real integration. All three proxies for financial development show a significant relationship with stock market returns. In line with what previous empirical evidence suggests, the sign of the coefficient is negative which implies that a country with a relatively more developed financial system responds more strongly to monetary policy surprises by the Fed and the ECB than a country where the financial system is relatively less developed, thereby sup-porting the hypothesis (H3a). In a similar research paper, Hausman and Wongswan (2006)

Table V: Cross-Country Results

This table shows estimates from the regression of Eurozone stock market returns on the surprise components of monetary policy actions and a proxy for financial development, financial integration, or real integration following:

Rt = α + β1 Ut + β2 (Ut * Xt)+ εt (5)

The sample consists of monetary policy data from January 1999 through December 2007, excluding the case of September 11, 2001. Financial Integration data only available from 2001 through 2006, hence contains 50 observations for the U.S. sample, and 73 observations for the Eurozone sample. Ireland and Luxembourg are excluded from the Financial integration sample due to abnormal values. Significant coefficients at the 5% level are in bold. Significance is below the coefficients in parentheses.

β1 β2 R2 N

Financial Development

Stock Market Capitalization / GDP -.049

(.194)

-.097*

(.011) .020 2178

Total Value Traded / GDP .006

(.867)

-.173**

(.000) .028 2178

Total Value Traded / Stock Market Capitalization -.017 (.657) -.135** (.000) .022 2178 Financial Integration

U.S. Equity Participation -.041

(.574)

-.272**

(.000) .094 500

Eurozone Equity Participation -.099

(.196)

.064

(.404) .003 730

Real Integration

Trade with U.S. / GDP -.356

**

(.000)

.161**

(.002) .065 872

Trade with Eurozone / GDP) .004

(.950)

-.034

(.570) .001 1306

** Significant at the .01 level. * Significant at the .05 level.

and the magnitude of stock market response. Although the explanatory power of their model was somewhat higher (4.4%), it still is relatively low and comparable to that in table V.

ship with Eurozone stock market returns. The sign of the U.S equity participation coeffi-cient is negative which implies that countries where U.S. equity participation is higher show a stronger response to U.S. monetary policy surprises. Alternatively, Eurozone stock market returns are not significantly related to Eurozone equity participation. This can be explained by the insignificant response of Eurozone stock markets to ECB monetary policy (see Tabel IV). When the response is insignificant it is even harder to explain dispersion in this response. Therefore, the corresponding hypothesis (H3b) is supported.

As discussed earlier in this paper, a more open financial integration leads to a higher mag-nitude of the reaction to U.S. monetary policy as financial investors will rebalance their portfolios after changes in macroeconomic variables. Hence, stock market indices where foreign equity participation is low will be focused on domestic monetary policy and less affected by changes in foreign monetary policy. Alternatively, more open economies with a relatively high percentage of equity participation will be more dependent on foreign in-vestors and foreign monetary policy. This theoretical relationship has been supported by empirical findings in recent research papers. In a research including 15 countries, Wongswan (2006) found financial integration to have a significant influence on the magni-tude of stock markets’ reaction to U.S. monetary policy, later confirmed by Ehrmann and Fratzscher (2006) and Hausman and Wonsgwan (2006) in more comprehensive researches. More specifically, Hausman and Wongswan (2006) found that stock markets with a rela-tively high level of financial integration with the U.S. respond more strongly to FOMC announcements. So, the positive relationship between financial integration and Eurozone stock market returns found in this research paper is consistent with both theory and find-ings in previous research papers.

Haus-man and Wongswan (2006) also found trade flows with the U.S. to be weakly significant, but found trade with the rest of the world to be insignificant. The striking difference be-tween the results from these two papers and the finding in this research paper, is that trade flows are negatively related with the reaction of Eurozone stock markets to U.S. monetary policy surprises, while both theory and previous research suggest the reverse. So, the find-ing in this paper suggests that countries with lower trade flows with the U.S. show a stronger response to U.S. monetary policy than countries that have a higher level openness. Therefore, the corresponding hypothesis (H3c) is rejected which suggested a positive

rela-tionship.

In sum, the findings related to the hypotheses are shown below in table VI.

Table VI: Results of the Hypotheses

Hypothesis Result

H1: Eurozone stock market returns are not related to raw changes

in Fed and ECB monetary policy instruments.

Supported

H2a: Eurozone stock market returns are negatively related to the

unexpected component of Fed and ECB monetary policy announcements.

Supported

H2b: Eurozone stock market returns are not related to the expected

component of Fed and ECB monetary policy announcements.

Not supported

H3a: The magnitude of the response of Eurozone stock markets to

monetary policy surprises is positively related with the degree of financial development.

Supported

H3b: The magnitude of the response of Eurozone stock markets to

monetary policy surprises is positively related with the degree of financial integration.

Supported

H3c: The magnitude of the response of Eurozone stock markets to

monetary policy surprises is positively related with the degree of real integration.

Not supported*

4.3 Robustness of the Results

The results presented above may be the result of model imperfection or the use of poor data. Therefore, three types of robustness checks are performed to test whether these re-sults are influenced by certain characteristics of the monetary policy surprises, same-day monetary policy decisions by the Bank of England, or outliers in the data.

The first issue to test is whether the strength of stock market returns are related to three characteristics of the monetary policy surprise: 1) whether there was a change in the mone-tary policy instrument or not, 2) whether the change at a particular meeting was a direc-tional change or not, and 3) whether the monetary policy surprise was positive or negative. These characteristics can have a significant unwanted influence on the response of stock market indices to monetary policy announcements. These tests are performed with a dummy variable. For the first characteristic the dummy is labeled “1” when the Fed and ECB changed the target rate of its monetary policy instrument, and “0” when they left the monetary policy instrument unchanged. For the second characteristic the dummy is labeled “1” when the change in the monetary policy instrument is a directional change, and “0” when the monetary policy instrument is unchanged or changed in the same direction as the previous change. For the third characteristic the dummy is labeled “1” when the monetary policy surprise (Ut) was positive, and “0” when negative. Table VII in appendix E shows

the results from regression equation 6 and the t-test. The results from the regression equa-tion show a significant relaequa-tionship of the stock markets to monetary policy surprises in the cases where the dummy equals “1”, and show an insignificant relationship in the cases where the dummy equals “0”. However, the t-tests reject the null-hypotheses for all three characteristics that the two coefficients are equal. This implies that these three characteris-tics of monetary policy surprises do not influence Eurozone stock market returns.

The final robustness test removes extreme values in Eurozone stock market returns from the sample, as they may have a have a significant influence on the results presented in this research paper. Extreme values are considered values 2.5 times the standard deviation from the regression estimate. The cases which are excluded from the analysis are given in Table VIII in appendix E, along with the results from the re-run regression equations 1 and 4. Panel A shows the baseline results excluding the extreme values. Excluding the extreme values, the sample with Fed monetary policy decisions only shows a significant response by Eurozone stock markets to raw changes in the federal funds rate. Moreover, the full sample which contains monetary policy decisions by both the Fed and the ECB is signifi-cant without the outliers. For both samples the R2 also improved significantly. Panel B in Table VII shows the regression results without outliers of regression equation 4. These results do not differ substantially from the results including the outlier data. More specifi-cally, none of the coefficients have changed in significance; Eurozone stock markets re-spond significantly to the surprise component and the expected component of Fed mone-tary policy, and respond insignificantly to both components of ECB monemone-tary policy. However, the R2 improved slightly for all three samples. Conclusively, these tests which exclude extreme values in stock market returns from the samples show that the results in the research paper are robust.

5

C

This research paper has examined the relationship between Eurozone stock market indices and monetary policy by the Federal Reserve and the European Central Bank over the pe-riod January 1999 through December 2007. Hereby, it has extended existing literature by examining and comparing the relationship of Eurozone stock markets with domestic (ECB) and foreign (U.S.) monetary policy announcements. In particular, the relationship of Euro-zone stock markets with the surprise component of monetary policy announcements is im-portant as that is the piece of information that has not been anticipated for. In contrast to previous research papers which often used the federal funds futures rate to isolate the sur-prise element, this research paper has used LIBOR because of its quotation in both euro and U.S. dollar. The relationships are examined by analyzing data of thirteen Eurozone stock market indices from twelve countries, utilizing the event-study methodology. Subse-quent cross-country analysis has examined how the response of Eurozone stock market to Fed and ECB monetary policy relates to proxies of financial development, financial inte-gration, and real integration. In relation to previous research two new variables for finan-cial market development have been used which measure liquidity in the finanfinan-cial markets.

The results show that Eurozone stock market indices in general are insignificantly related to the raw changes in monetary policy by both the Fed and ECB, consistent with other re-cent research papers. When the raw changes in monetary policy decisions are divided into a surprise component and an expected component using LIBOR, Eurozone stock market indices are more strongly related with U.S. monetary policy than with its domestic (ECB) monetary policy. These findings have also been reported in previous research, and are in line with the hypothesized relationships. However, the hypothesis that Eurozone stock markets are not related to the expected component of monetary policy announcements has been rejected. The results show that Eurozone stock market indices are positively related with the expected component of Fed monetary policy announcements. From these findings it can be concluded that U.S. monetary policy plays a dominant role in Eurozone stock market returns, more than domestic monetary policy.

finan-cial market development and the magnitude of the stock market response to monetary pol-icy surprises. So, both the size and liquidity of financial markets are positively related to the magnitude of the response to monetary policy surprises, which is in line with the hy-pothesized relationship. The proxy for financial integration also showed the hyhy-pothesized relationship, but only for the sample which contains Fed monetary policy surprises. Re-markably, the proxy for real integration showed a significant relationship in the reverse direction of the hypothesis: lower trade flows with the U.S. are associated with a higher stock market response to U.S. monetary policy surprises.

With these findings the research questions formulated in the introduction can be answered. Eurozone stock market indices are significantly related to Fed monetary policy announce-ments, and insignificantly related to ECB monetary policy announcements. More specifi-cally, there is a negative relationship with the surprise component, and a positive relation-ship with the expected component of Fed monetary policy announcements. Asymmetries in the response of Eurozone stock market indices to monetary policy surprises are positively related with financial market development and financial integration, and negatively related with real integration.

ap-to estimate a common sap-tock market reaction ap-to information, and a specific idiosyncratic shock. Rigobon and Sack (2004) and Craine and Martin (2003) used such models in which all data is used, but presented similar results to event-studies. Furthermore, the event-study methodology in combination with daily data does not rule out influences on stock market prices other than monetary policy. Stock markets may respond to other pieces of informa-tion released within the event window of the monetary policy announcement. By using high frequency intraday data this problem could be diminished, however these data was not available for this research.

R

Bentzen, E., P.R. Hansen, A. Lunde, & A. Zebedee, 2004. The Greenspan effect on equity markets: an intraday examination of US monetary policy announcements, Mimeo, Co-penhagen Business School.

Bernanke, B.S., & K.N. Kuttner, 2005. What explains the stock market’s reaction to fed-eral reserve policy?, The Journal of Finance, Vol. 60, No. 3, p. 1221-1257.

Bomfim, A.N., 2003. Pre-announcement effects, news effects, and volatility: monetary policy and the stock market, Journal of Banking & Finance, Vol. 27, p. 133-151. Bomfim, A.N., & V.R. Reinhart, 2000. Making news: Financial market effects of federal

reserve disclosure practices, manuscript. Federal Reserve Board.

Bredin, D., C. Gavin, & G. O’Reilly, 2004. US monetary policy announcements and Irish stock market volatility, Applied Financial Economics, Vol. 15, p. 1243-1250.

Bredin, D., S. Hyde, D. Nitzsche, & G. O’Reilly, 2007. UK stock returns and the impact of domestic monetary policy shocks, Journal of Business Finance & Accounting, Vol. 34, No. 5 & 6, p. 872-888.

Brown, S.J., & J.B. Warner, 1980. Measuring security price performance, Journal of

Fi-nancial Economics, Vol. 8, p. 205-258.

Brown, S.J., & J.B. Warner, 1985. Using daily stock returns: the case of event studies,

Journal of Financial Economics, Vol. 14, p. 3-31.

Conover, C.M., G.R. Jensen, & R.R. Johnson, 1999. Monetary environments and interna-tional stock returns, Journal of Banking & Finance, Vol. 23, p. 1357-1381.

Craine, R., & V. Martin, 2003. Monetary policy shocks and security market responses, Mimeo, University of California-Berkeley.

Davig, T., & J.R. Gerlach, 2006. State-dependent stock market reactions to monetary pol-icy, International Journal of Central Banking, Vol. 2, No. 4, p. 65-83.

Ehrmann, M., & M. Fratzscher, 2004. Taking stock: monetary policy transmission to eq-uity markets, Journal of Money, Credit, and Banking, Vol. 36, No. 4, p. 719-737. Ehrmann, M., & M. Fratzscher, 2006. Global financial transmission of monetary policy

Gürkaynak, R.S., B.P. Sack, & E.T. Swanson, 2002. Market-based measures of monetary policy expectations, Working paper, Board of Governors of the Federal Reserve Sys-tem.

Gürkaynak, R.S., B.P. Sack, & E.T. Swanson, 2004. Do actions speak louder than words? The response of asset prices to monetary policy actions and statements, Working paper, Board of Governors of the Federal Reserve System.

Hausman, J., & J. Wongswan, 2006. Global asset prices and FOMC announcements, Fed-eral Reserve Bank International Finance Discussion Papers No. 886.

Honda, Y., & Y. Kuroki, 2006. Financial and capital markets’ responses to changes in the central bank’s target interest rate: the case of Japan, The Economic Journal, Vol. 116, p. 812-842.

Jensen, G.R., & R.R. Johnson, 1993. An examination of stock price reactions to discount rate changes under alternative monetary policy regimes, Quarterly Journal of Business

and Economics, Vol. 32, No. 2, p. 26-51.

Jensen, G.R., & R.R. Johnson, 1995. Discount rate changes and security returns in the U.S., 1962-1991, Journal of Banking & Finance, Vol. 19, p. 79-95.

Jensen, G.R., R.R. Johnson, & W.S. Bauman, 1997. Federal reserve monetary policy and industry stock returns, Journal of Business Finance & Accounting, Vol. 24, No. 5, p. 629-644.

Krueger, J.T., & K.N. Kuttner, 1996. The fed funds futures rate a predictor of federal re-serve policy, Journal of Futures Market, Vol. 16, p. 865-879.

MacKinlay, A.G., 1997. Event studies in economics and finance, Journal of Economic

Literature, Vol. 35, p. 13-39.

Madura, J., & O. Schnusenberg, 2000. Effect of federal reserve policies on bank equity returns, The Journal of Financial Research, Vol. 23, No. 4, p. 421-447.

Patelis, A.D., 1997. Stock return predictability and the role of monetary policy, The

Jour-nal of Finance, Vol. 52, No. 5, p. 1951-1972.

Pearce, D.K., & V.V. Roley, 1983. The reaction of stock prices to unanticipated changes in money: a note, The Journal of Finance, Vol. 38, No. 4, p. 1323-1333.

Rigobon, R., & B. Sack, 2004. The impact of monetary policy on asset prices, Journal of