TARGET INTEREST RATE NEWS EFFECTS ON THE ASIA PACIFIC FINANCIAL MARKETS

Do Quoc Tho Nguyen

School of Banking and Finance The University of New South Wales

A dissertation submitted to the University of New South Wales in fulfillment of the requirements for the degree of Doctor of Philosophy (PhD)

2009

ORIGINALITY STATEMENT

I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged.

Signed Date 09/09/2009..............

ii

ABSTRACT
This thesis is the first study that provides comprehensive empirical evidence on both the impacts of the target interest rate news from the Reserve Bank of Australia (RBA) on the Australian financial markets, and the spillover effects of the target interest rate news from the US Federal Reserves (Fed) and the European Central Bank (ECB) on the Asia Pacifics equity and currency markets. This thesis contributes to the current literature in several ways. First, while there is ample evidence in the literature suggesting that the markets would not react to what is already expected but will react to the news, the current literature on the RBAs target rate effects is still limited to the investigation of the overall announcement impact on the first moment of the Australian market return only. Therefore, this thesis firstly comprehensively investigates the impacts of the unexpected components of the RBAs target rate announcements (or news) on the first two moments of various segments of the Australian financial markets including interest rate changes, the Australian dollar and stock market returns. In so doing, this thesis contributes to the current literature on the impacts of domestic target interest rate news. Second, while the established literature seems to be missing a thorough investigation of the spillover effects of the Feds and the ECBs news on the Asia Pacific markets, this thesis provides comprehensive evidence on the spillover effects of the Feds and the ECBs target rate news on both the mean and volatility of the Asia Pacifics stock and currency returns. Furthermore, we not only document the presence of the news spillover effects but also highlight the incremental explanatory power of the target interest rate news in the presence of the indirect effects from the

iii

USs and euro areas markets to the Asia-Pacific markets. To this end, this thesis contributes to the literature on spillover effects of foreign target interest rate news. Third, while the literature is silent on how quickly the target interest rate news is absorbed in foreign markets, this thesis takes a step forward and breaks down the daily horizon into the overnight and the intraday horizons. In so doing, the thesis examines the absorption speed of target rate news in the Asia-Pacific markets. This is an important issue because there might be potential for a diverse array of response dynamics across countries due to heterogeneous market developments, nature of monetary policy synchronization, and financial and real integration with the U.S. and the euro area. Specifically, this thesis presents three independent empirical inquiries that contribute to the literature on domestic and spillover effects of the target interest rate news. Chapter 41 provides comprehensive empirical evidence for the impacts of the RBAs target rate news on various segments of the Australian financial markets during the period from 1998 to 2006. We also investigate the spillover effects of the US Feds news on the Australian financial markets. We show that the RBAs and the Feds news significantly affect the Australian financial markets in line with a priori expectations. However, while the RBAs news raises volatility in the Australian financial markets, the volatility was significantly lower in all market segments following the Feds news.

A shorter version of this chapter is published in Research in International Business and Finance,

22(3), pp. 378-395, 2008. It has also received the Best paper award at the 2008 International Business and Economics Research conference, Nevada, USA.

iv

The spillover effects of the US Feds and the ECBs target interest rate news on the mean and the volatility of twelve Asia Pacifics stock markets returns are examined in Chapter 52, and seven Asia Pacific exchange rates against the US dollar and the euro over the period 1999-2006 are carried out in Chapter 63. The spillover effects on the conditional mean are generally consistent with the literature where a majority of Asia Pacific stock markets shows significant negative returns and a majority of currencies depreciates against the US dollar and the euro in response to the Feds and the ECBs unexpected rate rises. Furthermore, in response to the two target rate news, the conditional volatility of the Asia Pacific stock markets was higher while the market calming effects have been observed for the currency markets and both the Fed and the ECB news elicit persisting volatility responses. We conjecture that as the ECBs news tends to confirm the Feds earlier decision, this relationship might help reduce uncertainties in the Asia Pacific currency markets upon the future path of target interest rates from both the Central Banks, which ultimately results in into a lower volatility level. These findings are important not only to the Asia Pacifics policy makers to help them improve the conduct of monetary policy but also to market participants in designing trading mechanisms as well as risk management strategies in response to both domestic and external interest rate shocks. Furthermore, these findings also shed

A shorter version of this chapter is published in Journal of International Financial Markets,

Institutions & Money 19 (3), pp. 415-431, 2009.

3

A shorter version of this chapter is published in International Research Journal of Finance and

Economics, 20, pp. 27-45, 2008.

light on the lead-lag relationship between the Fed and the ECB in making policy decisions. The notion that the ECB follows the Fed in setting its policy is so strong amongst market participants that empirical evidence seems to be crucial. Despite the fact that the ECBs news arrives after the Feds news, this study provides evidence that the ECBs news has its own merits in the Asia Pacific markets and helps resolve differences in beliefs among market participants.

vi

ACKNOWLEDGEMENTS
Although only my name appears on the cover of this dissertation, I could never have reached the heights or explored the depths without the help, support, guidance and efforts of many people. I owe my gratitude to all those people who have made this dissertation possible and because of whom my graduate experience has been one that I will cherish forever. I would like to gratefully thank my sole supervisor, Associate Professor SukJoong Kim for his guidance, understanding, patience, and most importantly, his coauthorship attitude instead of supervision during my graduate studies at the University of New South Wales. His mentorship was paramount in providing a wellrounded experience consistent with my long-term career goals. I am also grateful to the guidance and endless support from Associate Professor Toan Pham as well as fellow research students during my study at the University of New South Wales. My special thanks also go to my colleagues at the International Monetary Fund, Raghuram G. Rajan (now at the University of Chicago), Stijn Claessens, Hali Edison, Susan Adams as well as Carl Hubbard at Trinity University for valuable comments and suggestions that greatly improve this thesis works. I would like to thank the examiners, Colm Kearney, Michael McKenzie and Eliza Wu, for valuable comments and suggestions that greatly enhance the thesis. I also would like to thank Stephen Ferris, Ali Fatemi, Ruggero Bertelli and Craig Lewis for giving me opportunities to referee papers submitted to your journals and Annual Meetings. I also would like to thank AusAID for generously providing me this second scholarship to pursue this PhD study. I also

vii

acknowledge financial support from the University of New South Wales for conference grants. Finally, and most importantly, I would like to thank my wife Ha Phi. Her support, encouragement, quiet patience and unwavering love were undeniably the bedrock upon which the past seven years of my life have been built. Her tolerance of my occasional vulgar moods is a testament in itself of her unyielding devotion and love. I thank my son, Ty, for his companion everyday to the University and his Childcare center as well as sharing with me his childhood overseas. I am also grateful to my parents and Has parents for their faith in me. This thesis is dedicated to my family!

viii

TABLE OF CONTENTS
Page ABSTRACT ................................................................................................................... iii ACKNOWLEDGEMENTS ......................................................................................... vii LIST OF TABLES........................................................................................................ xii LIST OF FIGURES..................................................................................................... xiv LIST OF ABBREVIATIONS .......................................................................................xv LIST OF JOURNAL PUBLICATIONS FROM THE DISSERTATION ............ xviii CHAPTER 1 - INTRODUCTION .................................................................................1 1.1. 1.2. 1.3. Objectives of the thesis ........................................................................................2 Contributions to the literature ...........................................................................4 Thesis outline .....................................................................................................10

CHAPTER 2 - LITERATURE REVIEW OF TARGET RATE NEWS EFFECTS ............................................................................................................12 2.1. The impacts of the central bank target interest rate on domestic financial markets ...........................................................................................................13 2.1.1. The debt markets .............................................................................................13 2.1.2. Stock markets ..................................................................................................15 2.1.3. Exchange rates .................................................................................................18 2.1.4. The credit channel ...........................................................................................19 2.2. 3.1. The spillover effects of foreign target interest rate news ...............................20 Central banks target interest rate communications .....................................25 CHAPTER 3 - DATA AND ECONOMETRIC MODELING ISSUES....................24 3.1.1. The RBAs cash rate target..............................................................................25 3.1.2. The U.S. Federal Reserves federal funds target rate ......................................27 3.1.3. The European Central Banks main refinancing rate target ............................28 3.2. Target interest rate news ..................................................................................30 3.2.1. The RBAs cash rate news ..............................................................................31 3.2.2. The Feds and the ECBs target rate news ......................................................32 3.3. Econometric methodology ................................................................................37 3.3.1. The RBAs effects on the Australians financial markets ...............................39 3.3.2. The Feds spillover effects on the Australian financial markets .....................40 3.3.3. The Feds and the ECBs spillover effects on the Asia Pacific financial markets ........................................................................................................................41

ix

CHAPTER 4 - THE REACTION OF THE AUSTRALIAN FINANCIAL MARKETS TO THE INTEREST RATE NEWS FROM THE RESERVE BANK OF AUSTRALIA AND THE US FED ............................................................46 4.1. 4.2. Introduction .......................................................................................................47 Data .....................................................................................................................50

4.2.1. Target interest rate announcement data ...........................................................50 4.2.2. Target interest rate news ..................................................................................52 4.2.3. Australian financial market daily returns ........................................................55 4.3. Empirical modeling issues ................................................................................56 4.3.1. Baseline EGARCH(1,1) model .......................................................................56 4.3.2. Overall news effects of the RBAs target interest rate news ...........................57 4.3.3. Asymmetric effects of the RBAs target rate news .........................................60 4.3.4. Spillover effects of the US Feds target interest rate news .............................61 4.4. Empirical results................................................................................................63 4.4.1. The baseline EGARCH(1,1) results ................................................................63 4.4.2. Overall effect of the RBAs target interest rate news......................................67 4.4.3. Asymmetric impacts of the RBAs target interest rate news ..........................71 4.4.4. Spillover effects of the US Feds target interest rate news .............................73 4.4.5. Asymmetric spillover effects of the US Feds target interest rate news .........75 4.5. 4.6. No-news effects ..................................................................................................75 Conclusion ..........................................................................................................76

CHAPTER 5 - THE SPILLOVER EFFECTS OF TARGET INTEREST RATE NEWS FROM THE US FED AND THE EUROPEAN CENTRAL BANK ON THE ASIA PACIFIC STOCK MARKETS ................................................................85 5.1. 5.2. Introduction .......................................................................................................86 Data and empirical modeling issues.................................................................89

5.2.1. The Asia Pacific stock index returns ...............................................................89 5.2.2. Empirical modelling ........................................................................................94 5.3. Empirical results................................................................................................99 5.3.1. Baseline model results .....................................................................................99 5.3.2. The US Feds target interest rate news spillover effects ...............................100 5.3.3. The ECBs target interest rate news spillover effects ...................................107 5.4. Robustness checks ...........................................................................................109 5.4.1. Joint spillover effect of the target interest rate news .....................................109 5.4.2. The role of indirect impacts...........................................................................111 5.4.3. Policy cycle impacts ......................................................................................120
x

5.4.4. Asymmetric effects ........................................................................................121 5.5. Conclusion ........................................................................................................129 CHAPTER 6 - TARGET INTEREST RATE NEWS EFFECTS ON THE ASIA PACIFIC CURRENCY MARKETS .........................................................................131 6.1. 6.2. 6.3. Introduction .....................................................................................................132 Data ...................................................................................................................134 Empirical methodology ...................................................................................136

6.3.1. Baseline EGARCH(1,1) model .....................................................................136 6.3.2. Target interest rate news spillover effects .....................................................139 6.4. Empirical results..............................................................................................141 6.4.1. Baseline models results ................................................................................141 6.4.2. The Feds target interest rate news spillover effects .....................................146 6.4.3. The ECBs target interest rate news spillover effects ...................................148 6.5. Robustness checks ...........................................................................................149 6.5.1. Joint spillover effects of the target interest rate news ...................................149 6.5.2. Asymmetric effects ........................................................................................151 6.5.3. The Feds and the ECBs news effect on the USDEUR pair ........................158 6.6. Conclusion ........................................................................................................159 CHAPTER 7 - CONCLUSION AND DIRECTIONS FOR FURTHER RESEARCH ..........................................................................................................161 7.1. 7.2. Conclusion ........................................................................................................162 Directions for future research ........................................................................166

7.2.1. Higher frequency data ...................................................................................166 7.2.2. Sectoral and firm level effects .......................................................................166 7.2.3. Determinants of the spillover effects .............................................................167 7.2.4. Other macroeconomic news spillover impacts ..............................................168 APPENDICES .............................................................................................................169 REFERENCES ............................................................................................................177

xi

LIST OF TABLES
Page Table 3.1- Descriptive statistics of the Feds and the ECBs target rate announcements and news ........................................................................................... 36 Table 4.1 - Descriptive statistics of the RBAs and the Feds target rate news ......... 51 Table 4.2 - Wald test of the unbiasedness of expectations of monetary policy announcements ........................................................................................................... 54 Table 4.3 - Descriptive statistics of the Australian financial market returns ............. 56 Table 4.4 - A priori expectations ................................................................................ 59 Table 4.5 - EGARCH(1,1) estimations of daily returns ............................................. 65 Table 4.6 - The RBAs news effects on the Australian financial markets ................. 70 Table 4.7 - The RBAs asymmetric surprises effects on the Australian financial markets ....................................................................................................................... 72 Table 4.8 - Spillover effects of the Feds target rate news in the Australian financial markets ....................................................................................................................... 79 Table 4.9 - Asymmetric spillover effects of the Feds target rate news in the Australian financial markets ....................................................................................... 80 Table 4.10 - Nonews effects on the Australian financial markets .............................. 81 Table 5.1 - Descriptive statistics of the Asia Pacific stock markets returns ............. 92 Table 5.2 - Granger causality tests for the Fed and the ECB ..................................... 98 Table 5.3 - Baseline model results ........................................................................... 101 Table 5.4 - The Feds target interest rate news spillover effects .............................. 106 Table 5.5 - The ECBs target interest rate news spillover effects ............................ 107 Table 5.6 - The US Feds and the ECBs Joint spillover effects .............................. 110 Table 5.7 - The Feds and the ECBs interest rate news effects estimated with indirect effects via index return spillovers ............................................................... 112 Table 5.8 - The Feds and the ECBs interest rate news effects on domestic markets .................................................................................................................................. 114 Table 5.9 - Spillover effects of the USs and the euro areas stock markets on the Asia Pacific markets ................................................................................................. 115 Table 5.10 - Marginal contribution of direct effect given indirect effect ................. 117 Table 5.11 - The policy cycle impacts ..................................................................... 122 Table 5.12 - The Fed and the ECB asymmetric effects............................................ 125 Table 6.1 - Descriptive statistics of the Asia Pacific currency markets returns...... 137 Table 6.2 - EGARCH(1,1) estimations of daily returns ........................................... 142 Table 6.3 - The Feds and the ECBs target interest rate news spillover effects ..... 147
xii

Table 6.4 - Joint spillover effects of the US Feds and the ECBs target interest rate news .......................................................................................................................... 153 Table 6.5 - Asymmetric news spillover effects ........................................................ 155 Table 6.6 - The Feds and the ECBs news effects on the USD/EUR pair .............. 158 APPENDICES Appendix A - Macroeconomic announcements made on the same date as the RBA's target rate announcements ........................................................................................ 170 Appendix B - Granger causality tests for the Fed and the RBA ............................. 171 Appendix C, Table C.1 - The US financial market returns and the US Feds target interest rate news effects .......................................................................................... 174 Appendix D - Controlled macroeconomic announcements ..................................... 176

xiii

LIST OF FIGURES
Page Figure 5.1 - Testing horizons...................................................................................... 91 Figure 5.2 - Direct and indirect effects ..................................................................... 111

xiv

LIST OF ABBREVIATIONS
ABS ANZ AUD CAB CAC Australian Bureau of Statistics Australia and New Zealand Banking Corporation Australian Dollar Current account balance French stock market index, Cotation Assiste en Continu (Continuous Assisted Quotation) CBA CBOT CET CPI DAX DJIA ECB ESCB EGARCH EST EUR Fed FOMC GARCH GDP HICP HSI IAPT Commonwealth Bank of Australia Chicago Board of Trade Central European Time Consumer Price Index German stock market index (Deutscher Aktien Index) Dow Jones Industrial Average Index European Central Bank European System of Central Banks Exponential autoregressive conditional heteroscedasticity Eastern standard time The Euro The United States Federal Reserve System Federal Open Market Committee Autoregressive conditional heteroscedasticity Gross Domestic Product Harmonised Index of Consumer Prices Hang Seng Index International Asset Pricing Theory
xv

IDR JIC JPY KLCI KRW KRX100 MPS

Indonesian Rupiahs Jakarta Stock Exchange Composite Index Japanese Yen Kuala Lumpur Stock Exchange Composite Index Korean Won Korea Exchange 100 Index MIT/Penn/Social Science Research Council (Modigliani (1971)s model)

MRO NAB NEER NKY NZD NZSE50FG

Main refinancing operations National Australia Bank Nominal effective exchange rate Japanese NIKKEI 225 Index New Zealand Dollar New Zealand Exchange Limited 50 Free Float Total Return Index

OMOs PCOMP PHP RBA RET RTT S&P/ASX 200 S&P500 SET SHCOMP

Open Market Operations Philippine Stock Exchange PSEi Index The Pilipino Peso (PHP) The Reserve Bank of Australia Retail sales growth rate Retail trade Standard & Poors/Australian Stock Exchange 200 Index Standard & Poor's 500 Index Stock Exchange of Thai Index Shanghai Stock Exchange Composite Index

xvi

STI TD TWD TWSE UE USD WBC

Singapores Straits Times Index Trade deficit Taiwanese Dollar Taiwan TWSE Index Unemployment rate The United States Dollar Westpac Banking Corporation

xvii

LIST OF JOURNAL PUBLICATIONS FROM THE DISSERTATION

1. A shorter version of Chapter 4 is published in Research in International Business and Finance 22 (3), pp. 378-395, 2008. It has also received the Best paper award at the 2008 International Business and Economics Research Conference, Nevada, USA. 2. A shorter version of Chapter 5 is published in Journal of International Financial Markets, Institutions & Money 19 (3), pp. 415-431, 2009. 3. A shorter version of Chapter 6 is published in International Research Journal of Finance and Economics 20, pp. 27-45, 2008.

xviii

CHAPTER 1 - INTRODUCTION

"Policymakers often have to act, or choose not to act, even though we may not fully understand the full range of possible outcomes, let alone each possible outcome's likelihood." Alan Greenspan, January 3, 2004

1.1. Objectives of the thesis Since the abandonment of monetary aggregate targeting in the mid-1980s, central banks of advanced countries have moved to targeting policy interest rates and public announcement of decisions on the interest rate target is a common practice in most central banks. Any change in the target rate constitutes an adjustment in the monetary policy stance, and the unexpected component of the change constitutes news. As the central banks target interest rate decisions impose interest rate risk to most market participants, it is widely documented that in the days approaching a central banks monetary policy meeting, market participants anticipate at least two important factors. First, they will predict the direction of the target rate action, i.e. whether the central bank will raise, cut or leave its target interest rate unchanged. Second, they will also anticipate the magnitude of such policy action, i.e. by how much the central bank will raise or cut its target rate? Accordingly, market participants must take positions based upon their expectations of the impending announcements of the central banks target interest rate stance. This expected part is, thus, already factored into the market prices observed immediately prior to the announcement. If the actual target rate announced is different from that already priced, markets react to this surprise (or news) component accordingly. Therefore,

central banks influence financial markets through the control over the target interest rates and the markets expectation on the future courses of the respective target rates. Furthermore, global financial markets have become increasingly integrated over the past decades. During this integration process, monetary policies play a key role in driving financial market linkages across national markets. Given the fast pace of real and financial integration, it is widely recognised that target interest rate decisions made by a central bank, especially policy decisions from the Fed, are keenly monitored by not only the domestic market participants but also by foreign market participants. However, the nature and the transmission channels through which target interest rate news dissipates are still not well understood. In addition, while the impact of the Feds monetary policy on domestic markets is well documented for the US markets, the spillover impact of the Feds target rate news to foreign markets is much less obvious. This thesis empirically investigates the direct impacts of the target interest rate news on the domestic financial markets, and the spillover effects from one central bank to foreign financial markets. Specifically, this thesis presents three empirical investigations that address the following important questions: 1. What are the nature, direction and magnitude of the impacts of the RBAs target interest rate surprises on the Australian financial markets? 2. What are the nature, direction and magnitude of the spillover effects of the Feds and the ECBs target rate news on the Asia Pacific financial markets?

3. The Feds and the ECBs target rate news would have both direct and indirect influences on foreign markets. Moreover, there is ample evidence that the Asia Pacific markets are influenced by the USs and the euro areas markets (the indirect effects). Therefore, given the presence of such indirect effects, whether the Feds and the ECBs target rate news has any incremental explanatory power?; and 4. How quickly is the target interest rate news from the Fed and the ECB absorbed in the Asia Pacific markets? 1.2. Contributions to the literature This thesis is conceptually linked to two strands of the literature on the monetary policy news. First, it contributes to the literature on the impacts of domestic target rate news by empirically investigating the impacts of the RBAs target rate news on the Australian financial markets via different channels ranging from interest rates, stock markets, exchange rates, and credit channels (via bank lending mechanism). Australian financial markets have been chosen as a case study in Chapter 4 due to its growing importance and integration with other markets in the Asia Pacific region and with the US. Furthermore, Australian financial markets are regarded as efficient markets as Edey and Elliott (1988) found, thus market responses to interest rate news could have been detected relatively quickly. Therefore, an empirical test on the spillover impacts of the Feds target interest rate news on the entire financial markets of a single country seems to be crucial before going further to the extensive and detailed investigations for such effects on the stock markets and the exchange markets of a large number of countries in the Asia-Pacific region in Chapters 5 and 6. However, the current literature on the RBAs cash rate effect is

limited to the investigation of the announcement impact on the first moment of Australian market returns, see for example Gasbarro and Monroe (2004), and Diggle and Brooks (2007). Therefore, a thorough investigation focusing on the impact of the unexpected component (or news) of the RBAs target rate announcements would provide direct and comprehensive evidence on the impacts of the RBAs target rate news on the Australian financial markets. Second, this thesis attempts to investigate the nature of the international spillover effects of a central banks target rate news on foreign markets. While the spillover impacts of the Feds news on some national markets have been documented, 4 the literature is missing a thorough investigation of the spillover effects of the Feds news on the Asia Pacific markets. This is a significant oversight as the information leadership role of the US in the Asia Pacific region is well documented, see inter alia Arshanapalli et al. (1995), Liu et al. (1998), Liu and Pan (1997), Ghosh et al. (1999), Ng (2000), Miyakoshi (2003), Kim (2003, 2005), Phylaktis and Ravazzolo (2005).5 In addition, while there are some studies focusing

See for example, Bredin et al. (2005) examine the Feds news effects on the Irish stock market

volatility; Ehrmann and Fratzscher (2003, 2005a) investigate the Feds news effects on the euro area money market returns, while Hausman and Wongswan (2006) examine such spillover effect on the stock, debt and foreign exchange markets. The Feds interest rate news has shown to be transmitted to these markets and the spillover effects are strongly felt. However, the transmission in the opposite direction is found to be weak, see Ehrmann and Fratzscher (2005a).
5

For Australia in particular, Kim and Sheen (2000) show that the Australian interest rates (90-day and

10-year rates) react strongly to the first and second moments of the corresponding US rate movements. Masih and Winduss (2006) report a straightforward cointegration relationship between the Australian and the US interest rates, whereas Narayan and Smyth (2004) show similar evidence

on the first moment spillover effects of the Feds target interest rate news on foreign markets, very limited evidence has been shown for the volatility spillover effects of the Feds news. This significant gap needs to be addressed because the target rate news does not only affect the direction of market movements in foreign markets (first moment influence) but also influences the trading environment, and hence the level of volatility (second moment effect). For example, Faff et al. (2009) shows that conditional stock market return volatilities have contemporaneous and positive systematic exposures to global return volatilities. Furthermore, Kearney (2000) shows that the global equity markets volatility is caused mostly by volatility in the US markets. Therefore, to the extent that the US stock market responds significantly to the Feds news, and that the foreign markets look to the U.S markets for trading momentum, there is potential for the Feds news to be significantly priced in foreign markets, especially in the Asia Pacific region where both financial and real integration with the US is growing rapidly. Another oversight in the literature is the relative lack of investigation of the spillover effects of the ECBs news.6 Given the increasing prominence of the euro areas equity markets, and the increasing real and financial integration between the Asia Pacific region with the euro area,7 there is strong potential that the ECBs news

for the stock markets. Moreover, Kim (2005) reports a direct causal information flow from the US stock market to that of Australia.
6

Although there are some recent studies on the impacts of the ECBs news on European stock

markets, see e.g. Bohl et al. (2008), and Bredin et al. (2009).
7

The EU is the largest trading partner for Australia and China; second largest trading partner for

Hong Kong, New Zealand, Taiwan, and Thailand; the third for Indonesia, Japan, Malaysia, and

would have spillover impacts on the financial markets in the latter. Finally and importantly, the literature is silent on how quickly the news from a central bank is absorbed in foreign markets. Against this background, this thesis contributes to the current literature in several ways. First, it provides comprehensive evidence for the impacts of the RBAs target rate news on both the mean and volatility of the Australian interest rate changes, the Australian dollar and stock market returns. In so doing, this thesis contributes to the current literature on the impacts of domestic target rate news. Second, this thesis fills significant gaps in the literature by providing comprehensive evidence on the spillover effects of the Feds and the ECBs target rate news on both the mean and volatility of the Asia Pacific stock and currency market returns. Furthermore, the thesis not only documents the presence of the news effects but also highlights the incremental explanatory power of the target interest rate news in the presence of the indirect effects from the USs and euro areas stock markets. Third, in order to shed light on the adjustment speed of the Asia Pacific markets to the Feds and the ECBs news, this thesis investigates the relative speed of news absorption by breaking down the daily horizon into the overnight and the intraday horizon. Thus, this thesis examines the extent of persistence effects, if any, of the Feds and the ECBs news effects on the Asia-Pacific markets.

Singapore; and the fourth for Korea and the Philippines (see http://ec.europa.eu/trade/issues/bilateral/ data.htm).

This thesis presents three independent empirical inquiries on the impacts of central banks target interest rate news on both domestic and foreign markets. In particular, Chapter 48 provides comprehensive empirical evidence for the impacts of the RBAs target rate news on various segments of the Australian financial markets, namely debt, stock and foreign exchange markets. We show that the RBAs news had a significant impact on the first moment of market returns/changes in line with a priori expectations, and the conditional volatility in most of the markets was significantly higher following the news. Asymmetric news effect is also observed for the Australian interest rate changes where markets tended to respond more strongly to unexpected rate rises than rate cuts. Furthermore, while the US Feds news influences only the USD/AUD exchange rate, the Australian market volatility was significantly lower in all market segments following the Feds news. These findings are informative not only to the RBAs decision makers to improve its conduct of monetary policy but also to market participants in designing trading mechanisms as well as risk management strategies in response to both domestic and external interest rate shocks. Chapter 5 9 examines the spillover effects of the US Feds and the ECBs target interest rate news on the market returns and return volatilities of twelve stock markets in the Asia Pacific region over the period 1999-2006. In addition to close-to-

A shorter version of this chapter is published in Research in International Business and Finance,

22(3), pp. 378-395, 2008.

9

A shorter version of this chapter is published in Journal of International Financial Markets,

Institutions & Money 19 (3), pp. 415-431, 2009.

close return over market close on calendar day t-1 to the close on calendar day t (close-to-close), it is split into the overnight horizon over the close of day t-1 to the open on day t (close-to-open), and the intraday horizon over the open on day t to the close on day t (open-to-close). This enables us to capture the Asia Pacific stock markets first reaction to the news (overnight returns) and delayed reactions, if any, during the trading day in the Asia Pacific region (intraday returns). The news spillover effects on the returns are generally consistent with the literature where a majority of stock markets shows significant negative returns in response to unexpected rate rises. Whilst the results of the speed of adjustment for the Feds news are mixed across the markets, the ECB news was absorbed more slowly, in general. The return volatilities of the Asia Pacific stock markets were higher in response to the interest rate news from both sources. In addition, both the Fed and the ECB news elicited delayed or persisting volatility responses. These findings have important implications for the Asia Pacifics policy makers and market participants alike in anticipating and managing potential spillover effects from the Feds and the ECBs so as to have forward looking policy responses and hedging strategies. Chapter 610 examines the spillover impacts of the US Feds and the ECBs target interest rate news on the first two moments of seven Asia Pacific exchange rates against the US dollar and the euro over the period 1999-2006. The spillover effects on the mean are generally consistent with the literature where a majority of currencies depreciates against the USD and the EUR in response to unexpected rate

10

A shorter version of this chapter is published in International Research Journal of Finance and

Economics, 20, pp. 27-45, 2008.

rises. Both the Fed and the ECB news elicited delayed or persisting volatility responses. The ECBs news tends to confirm the Feds decision. This relationship tends to help reduce volatility in the Asia Pacific currency markets. This finding provides useful information for the Asia Pacific monetary authorities, especially those who are targeting exchange rates, in designing and implementing their own monetary policies. These are also useful to currency traders in designing forward looking trading strategies. Furthermore, these findings also shed lights on the leadlag relationship between the Feds and the ECBs decision markings. The belief that the ECB follows the Fed in setting its policy is so strong among market participants such that empirical evidence seems to be crucial. Despite the fact that the ECBs news arrives after the Feds news, this chapter provides evidence that the ECBs news has its own merit in the Asia Pacific currency markets and helps to resolve differences in belief among market participants. In short, this thesis is the first study that provides comprehensive evidence on the influence of the RBAs target interest rate news on the Australian financial markets, and the spillover effects of the target interest rate news from the Fed and the ECB on both the first two moments of the Asia Pacifics equity and currency markets returns. 1.3. Thesis outline This thesis consists of seven chapters. Chapter 2 provides a literature review of the target interest rate news effects. Chapter 3 discusses the methods of isolating the unexpected component of the target interest rate announcements, and then proceeds to discuss the key methodologies used for empirical investigations contained in the following three chapters. Chapter 4 empirically investigates the

10

domestic effects of the RBA as well as the spillover effects of the Fed on the first two moments of the Australian debt, stock and currency markets. Chapters 5 and 6 focus on the spillover influences of the target interest rate news from the Fed and the ECB on the first two moments of the Asia Pacific stock and currency market returns, respectively. Chapter 7 concludes the thesis and offers directions for future research in the area of target interest rate news effects.

11

CHAPTER 2 - LITERATURE REVIEW OF TARGET RATE NEWS EFFECTS

12

2.1. The impacts of the central bank target interest rate on domestic financial markets This section discusses the underlying theory and empirical evidence for the influences of the central bank target interest rate on various domestic financial markets including the debt, equity and currency markets, which are the focuses of this thesis investigations. 2.1.1. The debt markets As Mishkin (1996) pointed out, to stimulate the economy the central bank increases the money supply, lowering a short-term nominal target interest rate, e.g. the Fed funds rate, resulting in lower short-term real interest rates assuming that the central bank has the capability to conduct Open Market Operations (OMOs) in ways such that it can vary the money supply to move the target interest rate in the desired direction. The expectation hypothesis of the term structure suggests that lower (higher) short-term real interest rates could result in lower or higher long-term real interest rates depending on market expectations on the future paths of nominal rate and inflation. The changes in long-term interest rates then stimulate (depress) consumer and investor spending raising (lowering) Gross Domestic Product (GDP) ultimately. The effect of announcements of monetary policy changes on the entire yield curve has been exhaustively investigated. Taylor (1995) shows that there is strong empirical evidence for substantial interest rate effects on consumer and investment spending through the interest rate channel. The earliest study in monetary announcement effects is Cook and Hahn (1989), who first documented the

13

connection between the Fed funds rate and market rates and showed that over the 1974-1979 period changes in the Fed funds rate had a significant positive impact on the yields of US Treasury securities at all maturities. However, their results might partly be due to a period of less systematic and transparent target interest rate decision making. The effect of a monetary policy decision on long-term rates might be different from that on shorter-term rates. Changes at the long end of the yield curve can at least in part be attributed to the markets views on the central banks credibility or its capability to control inflation. Therefore, a tightening of monetary policy can be compatible with a reduction in long-term interest rates if markets perceive the tightening as a credible step by monetary authorities to reduce inflation in the long-run (see e.g. Thornton, 1998). On the other hand, other studies argue that the effects of news surprises at the short and medium maturities mainly reveal information about market participants beliefs of the central banks reaction function (see e.g. Haldane and Read, 2000). In recent decades, as the market anticipation of monetary policy has improved, some of the potential impact of a target rate change would have already been incorporated into asset prices by the time the decision was made, causing the response to the actual announcement event to diminish (see Roley and Sellon, 1995). However, from technical perspectives, this might also be because of the failure to distinguish between the expected and unexpected components of the news. In his seminal work, Kuttner (2001) showed that Treasury yields did not respond to changes in the Fed funds rate that were already anticipated, while there was a large and significant impact of the unanticipated component of the interest rate

14

announcement. Demiralp and Jorda (2004) find that long-term rates react much more in unison during announcement days than at any other times. Recently, Valente (2008) shows that the Feds target rate news not only significantly affect the term structure of interest rates in the US but also spillover to Hong Kong and Singapore interest rates. On the volatility side, Lee (2006) empirically investigates the one-day response of interest rate volatility to Fed funds target rate changes over the period 1989-2003 and find a positive relation between the Feds news and the volatility at all maturities. The volatility response to an unanticipated Fed policy action is relatively large and highly significant for short-term interest rates. However, interest rates at long maturities are found to be responsive to target rate changes, even if they are anticipated when estimations take into account of structural change in association with the Fed's policy disclosure beginning in 1994. Notwithstanding there is strong empirical evidence for the Feds news impacts on interest rates at all maturities, Bernanke and Gertler (1995) pointed out that the macroeconomic response to policy-induced interest rate changes is considerably larger than that implied by conventional estimates of the interest elasticity of consumption and investment. This suggests that mechanisms other than the interest rates, e.g. asset prices including the stock prices and exchange rates, may also be at work in the transmission of monetary policy to the economy. 2.1.2. Stock markets In principle, monetary policy decisions can be transmitted to the economy via two channels of equity price movements: investment and wealth effects. Tobin's Theory of Investment provides a mechanism by means of monetary policy affecting
15

the economy through its influences on the valuation of equities (see Tobin, 1969). Tobin defines q as the market value of firms relative to the replacement cost of capital. If q is high, firms will invest more because adding capital is cheap as the market price of firms is high relative to the replacement cost of capital. As a result, investment spending will rise. On the other hand, when q is low, firms will not purchase new investment goods because the market value of firms is low relative to the cost of capital. How does monetary policy stance affect q? Expansionary monetary policy action can lead to a higher q in at least two ways. First, as lower interest rates leaving investors with less attractive alternatives or they have more money to spend. An attractive place to spend is in the stock market, thus, increasing demand for equities and consequently raising equity prices. Furthermore, lower interest rates making fixed income securities, e.g. bonds, less attractive relative to equities, thereby causing the price of equities to rise. In short, an expansionary policy stance will lead to higher equity prices, then a higher q, and thus higher investment spending. This channel has been empirically tested in various studies with mixed evidence. For example Blanchard et al. (1993) argue that as a result of fads, bubbles, or the influence of uninformed traders, stock prices may deviate from fundamentals. In such situations, managers may find it in their interests to ignore the signals arising from asset markets and invest on their own superior information. They find that given proxies for fundamentals market value appears to play a limited role in determining investment. Bond and Cummins (2000) make a similar argument and find that investment responds much more strongly to a measure of q constructed from profit forecasts than q constructed using the stock market. Gilchrist and Leahy

16

(2002) do not find a strong case for including asset prices in monetary policy rules. However, Ehrmann and Fratzscher (2004) show that individual stocks included in the S&P 500 react in a highly heterogeneous fashion and relate this heterogeneity to financial constraints and Tobin's q. In addition, Bjornland and Leitemo (2009) find great interdependence between the interest rate setting and real stock prices. Real stock prices immediately fall by seven to nine percent due to a monetary policy shock that raises the federal funds rate by 100 basis points. The second mechanism is via the wealth effect advocated by Modigliani (1971). In Modigliani (1971)s MPS model of consumption, the connection to monetary policy comes via the link between interest rates and asset prices. A policyinduced interest rate increase reduces the value of stocks and bonds, decreases households financial resources and leads to a fall in consumption and a lower level of real economic activity ultimately Previous studies consistently show that an unexpected target rate rise negatively affects the return of the US stock markets. For instance, Rigobon and Sack (2004) show that the US stock market decline in response to an unexpected increase in the Fed funds target rate derived from Eurodollar futures. Recently, Bernanke and Kuttner (2005) show that an unanticipated 25 basis point cut in the Fed funds target rate results in about a 1% increase in stock market indices. On the other hand, unexpected Fed funds target rate hikes have a positive impact on the volatility of the US stock markets. Consistent with this finding, Bomfim (2003) shows that the stock market has abnormally low conditional volatility on days preceding regularly scheduled policy announcements but significantly higher on the announcement day. Furthermore, higher than expected Fed funds target rate tends to have a larger effect

17

on volatility than negative surprises, which is consistent with both the leverage and volatility feedback hypotheses advocated by Black (1976) and French et al. (1987). 2.1.3. Exchange rates The monetary policy effects run from interest rates to exchange rates via the uncovered interest rate parity condition relating interest rate differentials to expected exchange rate movements. Thus, a cut in the target interest rate relative to foreign interest rates would lead to a weaker currency due to now relatively lower real domestic interest rate and would stimulate net exports and the overall level of aggregate demand. There is strong evidence for the impacts of monetary policy on exchange rates, such as Faust and Rogers (2003) and Faust et al. (2003) find delayed overshooting effects of the Feds news on the British Pound and German Mark/Euro, and such policy shocks can explain only a small portion of currency variability. Andersen et al. (2003) and Ehrmann and Fratzscher (2005b) show that the USs and the euro areas (and Germans news before 1999) macroeconomic news have a significant impact on the US dollar-euro exchange rate. Fatum and Scholnick (2008) find that the news component of a tightening (loosening) of US monetary policy is associated with a same day appreciation (depreciation) of the US dollar against the British Pound, German Mark, and Japanese Yen. Furthermore, Lobo et al. (2006) report that changes in the Feds interest rate target are positively related to changes in the value of the dollar, and also provide evidence for the asymmetric impact of the Feds news where unexpected rate rises have a larger effect than unexpected rate cuts for the Pound, Mark, and Canadian dollar, whereas the opposite is true for the Japanese Yen.
18

2.1.4. The credit channel Asset values also play an important role in the credit channel advocated by Bernanke and Gertler (1995). Advocates for the credit channel suggest that monetary policy affects not only the general level of interest rates but also the size of the external finance premium. Bernanke and Gertler (1995) suggested two mechanisms explaining the link between monetary policy actions and the external finance premium: the balance sheet channel and the bank-lending channel. In the balance sheet channel, asset prices are especially important in that they determine the value of the collateral that firms and consumers may present when obtaining a loan. In frictionless credit markets, a fall in the value of borrowers collateral will not affect investment decisions. However, in the presence of information or agency costs, declining collateral values will increase the premium borrowers must pay for external finance, which in turn will reduce consumption and investment. Thus, the impact of policy-induced changes in interest rates may be magnified through this financial accelerator effect. The bank-lending channel relies on credit market frictions where banks play a central role as they are especially well suited to solve asymmetric information problem in credit markets. The essential insight is that because banks rely on reservable demand deposits as an important source of funds, contractionary monetary policy, by reducing the aggregate volume of bank reserves, will reduce the availability of bank loans. In turn, as most firms and households rely heavily or exclusively on bank financing, a reduction in loan supply will ultimately depress aggregate spending.

19

The empirical evidence on the credit channel is somewhat mixed. For example, Miron et al. (1994) and Driscoll (2004) find little support for the banklending channel of monetary transmission. However, Kashyap et al. (1993), Kashyap and Stein (2000), Bernanke (1996), and Peersman and Smets (2005) provide supporting evidence for the bank-lending channel. The two studies that examine the monetary transmission by looking at the response of disaggregated stock returns to monetary shocks show opposite results. Warner and Georges (2001) find no evidence supporting a credit channel. In contrast, consistent with the credit channel, Ehrmann and Fratzscher (2004) document firm-level heterogeneity in the effect of monetary news on stocks based on financial constraints. 2.2. The spillover effects of foreign target interest rate news As the global financial market is increasingly integrated, it is necessary to understand the driving forces behind such integration. Among this, international transmissions or spillover effects of monetary policy actions of a central bank to foreign markets have attracted attention. Target rate news from major economies might affect other national markets via at least three channels. 11 Firstly, foreign target rate news would directly affect domestic markets if domestic monetary authorities were targeting exchange rates. An unexpected monetary policy stance in the US, for example, would results in a corresponding action from other countries, e.g. Singapore or Hong Kong, to maintain the exchange rate targets. Secondly, due to increasing global financial integration, monetary policy news from the US, for instance, might affect the markets in other countries via capital flows and other

11

See, for example Ehrman and Fratzscher (2003).

20

arbitrage activities. The final channel is through real integration such that foreign announcements might reveal important information about macroeconomic conditions of other countries. For example, if foreign monetary policy decisions change domestic macroeconomic conditions through effects on the terms of trade with the trading partners, such decisions may provide news not only about foreign economic conditions but also about the prospects of the domestic economy.12 The literature documents some evidence for the spillover impacts of the Feds funds rate news on foreign markets. For example, Ehrmann and Fratzscher (2003, 2005a) show that interest rates in the US, Germany and euro area react strongly to respective domestic target rate news. Furthermore, the response to foreign target rate news differs across markets where both German and euro area markets respond strongly to the Feds news, the opposite direction effect is weaker. Another strand of the target rate spillover effect literature is to identify the determinants of the spillover effects across markets. Ehrmann and Fratzscher (2006) focus on the link between the US monetary policy and 50 foreign stock markets and show that the US monetary policy news has significant and sizable effects on foreign stock prices. Moreover, they provide evidence that a countrys real and financial linkages with the world, and not the bilateral integration with the US, are key determinants of cross-country variation in response. In particular, they show that the sensitivity to US monetary policy news is higher in those countries that are more integrated globally rather than bilaterally with the US in real and financial terms. This suggests that domestic financial conditions in any country do not necessarily

12

See Ehrmann and Fratzscher (2005a) for further details.

21

depend on the financial conditions in the US market, but on some other effects that are due to the complex mechanisms of the global financial and real linkages. On currency markets, Fratzscher (2008) analyses the link between economic fundamentals, including the Feds news, and exchange rates from a cross sectional perspective. He shows that the USs shocks, including the Feds news, exerted heterogeneous impacts across 26 main currencies in the basket of the US dollar trade-weighted nominal effective exchange rate (NEER). More importantly, he shows that not only exchange rate regimes but also the degrees of financial integration are key determinant of such heterogeneity in responses. In a more comprehensive study, Hausman and Wongswan (2006) investigate the Feds news impact on the stock, debt and foreign exchange markets of 49 countries including both developed and emerging markets. They also take a step further to explore the transmission channels of the US monetary policy to foreign economies via real economic linkages, financial linkages, and exchange rate regimes. They find that foreign asset prices respond to the Feds news and cross-country responses are related to a countrys exchange rate regime. Equities and interest rates respond more strongly to the Feds news in countries with less flexible exchange rate regime, while exchange rates respond less. Financial integration, proxied by equity market capitalization owned by the US investors, and trade linkages with the US are also strongly related to cross country responses. In summary, this thesis is well related to two strands of the literature on the monetary policy news impacts as highlighted above. First, it provides comprehensive evidence on the target interest rate news on the key segments of the Australian financial markets, namely debt, stock and foreign exchange markets. Second, this

22

thesis documents and discusses evidence for the international spillover effects of the Feds and the ECBs target interest rate news on both the conditional mean and volatility of the Asia Pacific stock and exchange rate returns.

23

CHAPTER 3 - DATA AND ECONOMETRIC MODELING ISSUES

24

This chapter begins with discussions on the institutional arrangements of key policy interest rates targeted by three central banks considered in the later chapters. Then, Section 3.2 shows how the unexpected component of the target rates is generated and Section 3.3 discusses modeling issues. 3.1. Central banks target interest rate communications The last two decades have witnessed a surge in central banks efforts to improve policy transparency via various forms of policy communications to the public. This is because there is a strong notion that a prerequisite for an independent central bank is its accountability to the public. Therefore, central banks have responsibilities to timely inform policy actions to the public and explain to the extent possible the underlying rationales for such policy actions. Although central banks with similar objectives follow different methods of communication to the public, a common theme is that central banks would communicate their decisions on the target interest rate to the public. The timing and details of this communication however differ across central banks (see Blinder et al., 2008). The following sections briefly provide backgrounds of the three central banks target rates: the RBAs cash rate, the Fed funds rate and the ECBs main refinancing rate. 3.1.1. The RBAs cash rate target The Reserve Bank Act 1959, Section 10(2) states that It is the duty of the Reserve Bank Board, within the limits of its powers, to ensure that the monetary and banking policy of the Bank is directed to the greatest advantage of the people of Australia and that the powers of the Bank ... are exercised in such a manner as, in the opinion of the Reserve Bank Board, will best contribute to:

25

(a) the stability of the currency of Australia; (b) the maintenance of full employment in Australia; and (c) the economic prosperity and welfare of the people of Australia. In practice, the RBAs monetary policy's principal medium-term objective is to control inflation. In line with previous understandings between the Federal Government and the RBA in the Statement on the Conduct of Monetary Policy issued in 2007, the Governor and the Treasurer agreed that the appropriate target for monetary policy is to achieve an inflation rate of 2-3 per cent on average. Therefore, the inflation target is the core of the Australian monetary policy framework. It provides discipline for monetary policy decision-making, and serves as an anchor for private sector inflation expectations. The RBAs Board usually meets eleven times each year on the first Tuesday of the month except in January to decide on its monetary policy stance. Therefore, market participants know the meeting dates in advance. Through OMOs, by managing the supply of funds available to banks in the official money market, the RBAs Domestic Markets Department maintains conditions in the official money market to keep the cash rate, which is the rate charged on overnight loans between financial intermediaries, at or near an operating target decided by the Board. From January 1990, the RBA started to publicly announce its monetary policy decisions. From January 1998 until November 2007, the RBA Board's decision to change or leave the cash rate unchanged is announced in a media release at 9:30 am Australian EST (GMT+10) one day following the board meeting. Since December 2007, a media release is issued at 2.30 pm after each Board meeting.

26

3.1.2. The U.S. Federal Reserves federal funds target rate Federal Reserve Act, Section 2A states that the Board of Governors of the Federal Reserve System and the FOMC shall maintain long-run growth of the monetary and credit aggregates commensurate with the countrys long-run potential to increase production, to promote effectively the goals of maximum employment, stable prices and moderate long-term interest rates. However, it has been argued that despite its multiple objectives, the Fed has traditionally placed more emphasis on achieving price stability and, in recent years, there have been calls for a clearer price stability mandate for the Fed (see e.g. Wynne, 1999 and Bernanke, 2003). The major tool the Fed uses to influence the supply of reserves in the banking system is OMOs conducted at the Federal Reserve Bank of New York. The shortterm objective for OMOs is specified by the FOMC. This objective can be a desired quantity of reserves or a desired price (the Fed funds rate). 13 Although the Feds objective for OMOs has varied over the years, during the 1980s, the focus gradually shifted toward attaining a specified level of the Fed funds rate. Since 1994, the FOMC has been announcing the Fed funds target rate after its regularly scheduled (eight meetings a year) and ad hoc meetings at 2:00 pm US Eastern Standard Time (EST, GMT-5), unless otherwise specified. After each meeting, the Fed provides a short press release containing the decision, a concise explanation of its underlying reasoning and sometime some forward-looking hints.

13

The Fed funds rate is the interest rate at which depository institutions lend balances at the Fed to

other depository institutions overnight.

27

Following is a typical example of the FOMCs press release after its meeting on January 31, 2007. The Federal Open Market Committee decided today to keep its target for the federal funds rate at 5-1/4 percent. Recent indicators have suggested somewhat firmer economic growth, and some tentative signs of stabilization have appeared in the housing market. Overall, the economy seems likely to expand at a moderate pace over coming quarters. Readings on core inflation have improved modestly in recent months, and inflation pressures seem likely to moderate over time. However, the high level of resource utilization has the potential to sustain inflation pressures. The Committee judges that some inflation risks remain. The extent and timing of any additional firming that may be needed to address these risks will depend on the evolution of the outlook for both inflation and economic growth, as implied by incoming information. As can be seen, the first paragraph clearly and concisely announces the FOMCs target rate decision, i.e. to leave the Fed funds target rate unchanged. The next two paragraphs state the underlying rationales for the decision and the last paragraph offers some hints on the likelihood of the FOMCs decisions in the future meetings. 3.1.3. The European Central Banks main refinancing rate target Article 105 (1) of the Maastricht Treaty states that the primary objective of the European System of Central Banks (ESCB) shall be to maintain price stability and that without prejudice to the objective of price stability, the ESCB shall support
28

the general economic policies of the Community with a view to contributing to the achievement of the objectives of the Community as laid down in Article 2. In this respect, Article 2 states that the objectives of the Community are, inter alia, to ensure a high level of employment (), sustainable and non-inflationary growth, and a high degree of competitiveness and convergence of economic performance. As the ECB was established as the core of the Eurosystem and the ESCB, the Treaty thus establishes a clear hierarchy of objectives for the ECB and assigns overriding importance to price stability. Moreover, the ECB has made public its precise quantitative definition of price stability. The ECB has defined price stability as a year-on-year increase in the Harmonised Index of Consumer Prices (HICP) for the euro area of below 2%. In the pursuit of price stability, the ECB aims at maintaining inflation rates below, but close to, 2% over the medium term. For the ECB, the interest rate on the main refinancing operations (MRO), regular liquidity-providing reverse transactions with a frequency and maturity of one week, is perceived to be the target policy interest rate as it plays a pivotal role in pursuing the ECBs OMOs.14 Although the Governing Council meets twice a month, it normally makes a monetary policy decision at the first of the two meetings, after

14

The other key interest rates for the euro area are: (i) the rate on the deposit facility which banks can

use to make overnight deposits with the national central banks and normally provides a floor for the overnight market interest rate; and (ii) the rate on the marginal lending facility, which banks can use to obtain overnight liquidity from the national central banks against eligible assets, and this rate normally provides a ceiling on the overnight market interest rate. These two rates are for standing facilities, which aim to provide and absorb overnight liquidity, signal the general monetary policy stance and bound overnight market interest rates.

29

which a press release announcing the decision on the key ECB interest rates is made at 1:45 pm Central European Time (CET, GMT+1). Thus, the ECBs Governing Council has aligned its policy interest rate decision frequency closely with that of the Fed.15 However, unlike the other two central banks, the ECB does not only release a press statement with the policy decision but also holds a press conference on the Governing Councils meeting day, including a question and answer (Q&A) session at 2:30 pm CET. This thesis empirically investigates the impact of news arising from the ECBs press releases at 1:45 pm CET only. 3.2. Target interest rate news Empirical work that fails to decompose monetary policy changes into expected and unexpected components is also likely to lead to biased results due to errors in variables. In particular, a number of theories based on the assumption of efficient markets would suggest that only unanticipated changes in policy should influence asset prices immediately (i.e. on the announcement day of a monetary policy change asset prices should respond only to the surprise element of such a change). On the other hand, anticipated changes in policy should not affect asset prices but instead such information should have already been factored into the asset prices observed immediately prior to the announcement. Otherwise, arbitrage opportunities would exist and markets would be deemed inefficient. Studies that examine the influence of policy rate changes and fail to decompose actual changes

15

This is the case since November 8, 2001 as announced by the President of the ECB. However, the

Governing Council can still decide to change the key ECB interest rates at any time regardless of previously scheduled meetings if needed.

30

into these two components are liable to lead to biased results. For example, Cook and Hahn (1989) fail to take account of expected and unexpected changes in monetary policy and so their results are not credible. Other studies that suffer from this problem include Concover et al. (1999) and Durham (2001). Recent research has attempted to distinguish between unexpected and expected changes in target interest rate announcements. The news component is estimated by (i) the difference between the actual announcement and the median/mean of market surveys; (ii) derivation of unexpected component from future markets data; or (iii) derivation of expectations based on forecasts from regression analysis then subtract these expectations from the actual announcements for unexpected component. However, as liquidity and range of instruments offered in the futures markets have been improving, one can directly derive a measure of the surprise element on a continual basis using data on futures contracts. Therefore, the second approach has become popular and is also used in this thesis. 3.2.1. The RBAs cash rate news The best proxy for Australian market expectation on the RBAs upcoming target rate announcement is the 30-Day Interbank Cash Rate Futures contract. However, the launch date for this contract was August 12, 2003, thus it is not practical for this investigation. We instead use the market surveys of target interest rates from January 1998 to December 2006. Days surrounding the RBAs target rate announcements, the financial press reports what the market consensus was at the time of each announcement. These reports were searched on the Factiva database over a few days before and after the RBAs scheduled announcements. The unbiasedness and efficiency of the expectation data generated this way has been
31

tested and the results suggest that the market-based expectations are unbiased predictors of the RBAs interest rate announcements where the null hypothesis cannot be rejected even at 90% level of confidence. The news component of each announcement is then the difference between the actual target rate change announcement and the market expectation. Chapter 4 discusses this methodology of news extraction in details. 3.2.2. The Feds and the ECBs target rate news For the Fed, some of the earlier studies employed market survey expectations to proxy for expected target rate announcements (e.g. Reinhart and Simin, 1997). However, recent studies have instead relied on market price-based proxies thanks to the developments of the financial markets. Krueger and Kuttner (1996) find that the Fed funds futures rate is an efficient predictor of the Fed funds target rate, and therefore an appropriate market-based measure of policy expectations. Grkaynak et al. (2007) later confirm this finding. In his seminal work, Kuttner (2001) uses the Fed funds futures rates to separate the target rate changes into anticipated and unanticipated components. He finds that the US Treasury bill, note and bond yields responses to anticipated changes in the target rate are small, while the responses to unanticipated changes are large and significant. Bomfim (2003) extends Kuttner (2001) to asset return volatilities and finds that asset returns are more volatile following announcements containing unexpected rate changes. On the other hand, the early literature on the ECBs news employed pricebased proxies to gauge the market expectations on the ECBs target rate announcements. However, the choice of market instruments differs across researchers. Gaspar et al. (2001) use EONIA (Euro OverNight Index Average, the
32

effective overnight reference rate for the Euro) to gauge the probability attached to a change in the ECBs target interest rate before the Governing Councils meeting. Perez-Quiros and Sicilia (2002) propose a principal components approach that utilizes the daily changes of different money market interest rates including the EONIA, the one-week, one-, two- and three-month EONIA swap rates and the closest three-month EURIBOR futures rates. Their approach is to extract the key common component that shapes the evolution in all the above rates. Wrtz (2003) measures the interest rate change expectations from the forward rate implied by the one- and two-month EONIA swap rates. However, due to the high volatility and the impacts of liquidity considerations rather than the monetary policy considerations as identified by Bindseil (2002) in underbidding scenarios, it seems that the EONIA is not the best proxy for the market expectation on the ECBs upcoming interest rate announcements. Bernoth and Von Hagen (2004) find that the three-month EURIBOR futures rate is an unbiased predictor of the euro area policy rate changes. Thus, the literature seems to suggest that a market-based approach using futures rates would provide us with the markets unbiased expectations on the ECBs upcoming interest rate announcements. Ehrmann and Fratzscher (2003, 2005a) utilize the Reuterss survey of 25-30 market participants conducted on the Friday before each meeting of the ECBs Governing Council as a proxy for the market expectations on the upcoming interest rate decision. However, these surveys fail to capture any change in market expectations that occur between Friday and the actual announcement date. Against this background, this thesis uses Kuttner (2001)s methodology to generate the unexpected components of the Feds and the ECBs target rate announcements from January 1999 to December 2006. Current-month Fed funds futures contracts traded on the Chicago Board of Trade (CBOT) are used to extract
33

the Feds news, and the three-month EURIBOR futures contracts traded on the EUREX for the ECBs surprises.16 The news component of the Feds target rate announcement on day d of month m can be derived from the implied change in the price of the futures contract. Since the Fed funds futures settlement price is based on the monthly average of the spot Fed funds rate, it is necessary to account for the number of days affected by the announcement in that particular month as in equation (3.1). The 3-month EURIBOR futures settlement price is based on the reference interest rate (EURIBOR) for threemonth euro term-deposits on the last trading day, and so the news component of the ECBs target interest rate announcement is calculated as in equation (2.1) without the scaling factor D/(D-d).
i u = D 0 0 fm , d f m , d 1 Dd

(3.1)

where: i u is the unexpected target rate change;

f 0m, d is the current month Fed funds futures rate for the Fed and three-

month EURIBOR futures rate for the ECB;

f 0m, d-1 is the futures rate on the day prior to the announcement; D is the number of days in the month; and D-d is the number of days in the month affected by the announcement.

16

Although 1-month EONIA and EURIBOR Futures contracts are better proxies, the former was

introduced only on January 27, 2003, and the latter was delisted on March 16, 2004, hence they are not practical for this study. 34

The target interest rate data of the two central banks were obtained from their respective websites. 17 Panel A of Table 3.1 reports the breakdown of target rate announcements into rate rises, rate cuts and unchanged sub-components. From January 1999 to December 2006, the Fed and the ECB made 69 and 131 target rate announcements, respectively. Of these, the Fed had 19 announcements with rate changes (13 rate rises and 6 cuts) and 50 with no changes. The ECB made 21 announcements of the target rate changes (13 rises and 8 cuts) and 110 with no change. Thus, most of the interest rate announcements contained no change (72% for the Fed and 84% for the ECB). Panel B of Table 3.1 reports the summary statistics for the interest rate news series. While 35% of the Feds interest rate announcements were correctly anticipated, only 26% of the ECBs interest rate announcements were correctly expected. Whereas more negative surprises than positive surprises were observed for the ECB (44% compared with 31%), there were more positive than negative surprises for the Fed (36% compared to 29%). On average, the Feds and the ECBs target rate change announcements were lower than the markets expectation (-0.0639 and -0.0050 percent, respectively). The variance of the ECBs news is much lower than that of the Feds (0.0009 and 0.1027 respectively). Furthermore, the news series demonstrate strong evidence of negative skewness and leptokurtosis.

17

These data are available at http://www.federalreserve.gov/FOMC/#calendars for the Fed, and

http://www.ecb.int/stats/monetary/rates/html/index.en.html for the ECB.

35

Table 3.1- Descriptive statistics of the Feds and the ECBs target rate announcements and news

This table reports descriptive statistics for the Feds and the ECBs target interest rate announcements (Panel A), the surprise components (Panel B) for the period from January 1999 to December 2006.
ECB target rate announcements

Panel A. Target interest rate announcements Fed fund target rate announcements

No. of announcements Proportions

Total 69 (100%) Total 131 (100%)

Rate rise 13 (19%)

Rate cut 6 (9%)

No change 50 (72%)

Rate rise 13 (10%)

Rate cut 8 (6%)

No change 110 (84%)

Panel B. Target interest rate surprises

No. of observations Proportions Summary statistics Mean -0.0639 0.0994 -0.3447 -0.0050 0.0223 -0.0270 Variance 0.1027 0.0075 0.2349 0.0009 0.0006 0.0007 Skewness -3.5799 0.8090 -1.8831 -0.2408 3.9062 -2.1255 Excess Kurtosis 18.6141 2.1375 6.4368 10.3984 20.5803 7.3462 Min -1.9140 0.0055 -1.9140 -0.1300 0.0100 -0.1300 Max 0.2760 0.2760 -0.0050 0.1500 0.1500 -0.0100 Note: The numbers of surprises shown in Panel B are different from the actual rate change announcements for both central banks. This is because even when there was no rate change announcement, markets might have expected a rate rise or fall and so the surprise component is non-zero.

Total 69 (100%)

Fed fund rate surprises Positive Negative suprises surprises 25 20 (36%) (29%)

No surprise 24 (35%)

Total 131 (100%)

ECB policy surprises Positive Negative suprises surprises 40 57 (31%) (44%)

No surprise 34 (26%)

36

3.3. Econometric methodology As this thesis examines not only the target news impacts on the conditional mean but also on the conditional volatility, the chosen econometric model would firstly needs to be well suited to modeling daily financial returns series characterized as skewed, leptokurtic and non-normally distributed with time-varying second moments as documented in the literature. Furthermore, many studies such as Black (1976), Christie (1982), Nelson (1991), Glosten et al. (1993), and Engle and Ng (1993) find an asymmetric relation between stock returns and volatility. Specifically, they find that stock returns are negatively correlated with changes in returns volatility, i.e. lower than expected returns would lead to higher volatility and higher than expected returns tend to reduce volatility. However, GARCH models assume that only the magnitude and not the sign (positive or negative) of unexpected excess returns determine volatility features as they impose the assumption that the conditional volatility of the asset is systematically affected by positive and negative innovations (Nelson, 1991). Moreover, as the volatility effects of the news depend on the role of the news in resolving the heterogeneity of beliefs and expectations among market participants, the hypotheses tested in this thesis include not only the volatility increasing but also the volatility reducing (market calming) influence of the news. If the former dominates then one would expect to observe a positive news coefficient in the second moment and the non-negativity constraints of GARCH models would not be an issue.

37

However, if the market calming influence dominates, then there would be a negative relationship between the news and the conditional variance and so the news coefficient would be expected to be negative. In this case, the ubiquitous GARCH models fail to capture such impact, as they require the coefficients in the conditional variance equations to be all positive in order to satisfy the non-negativity constraints of the conditional variance. However, Nelson (1991)s EGARCH specifications allow both positive and negative coefficients and so are better suited to testing the hypotheses of the volatility raising and market calming effects of the interest rate news. In this thesis, we use the Nelson (1991)s EGARCH (1,1) models to simultaneously investigate the impacts of target rate news on both the conditional mean and volatility of asset prices. The investigations start with the baseline univariate EGARCH(1,1) model to model each of the return series, and then progress to specific model specifications to test the impacts of the target rate news. The baseline EGARCH (1,1) model can be described by the conditional mean and the conditional variance equations (3.2a) and (3.2b) shown below.
y t = c + Lag y t 1 + Mon Mon t + Hol Hol t + t

ln ht = c + h ln ht 1 + 1

t 1
ht 1

+ 2

t 1
ht 1

+ Mon Mont + Hol Hol t

(3.2a) (3.2b)

The conditional mean equation for the returns ( yt ) is expressed as a function of a one day lag of returns and a Mon and/or the Hol dummies in relevant markets. The Mon variable is used to account for the weekend effect and takes the value of one for Mondays and zero otherwise to account for the weekend effect. The Hol dummy is used to account for the potential impacts of differences in trading days across markets due to market closure. It takes the number of days of each market
38

closure between two successive market prices. For example, for normal consecutive daily observations (i.e. returns calculated over two days - Monday close to Tuesday close price) the value of zero is assigned, whereas values of one or higher will be assigned for returns observations calculated over a longer horizon due to market closure. This seasonal dummy helps to control for the days of more intense information flows following a longer period of market closure. The conditional variance equation for the returns in the financial market series ( ht ) is expressed as a function of one period lags of the variance and the residuals, and the Mon and/or Hol dummies. 3.3.1. The RBAs effects on the Australians financial markets In Chapter 4, to investigate the impacts of the RBAs target rate news on the Australian financial markets, we add a news variable, RBANewst to equations (3.2a) and (3.2b). In addition, in the course of a trading day, other macroeconomic announcements, apart from monetary policy stance, might hit the markets. We include a dummy variable (MacroAnni,t) for each of the Australian macroeconomic announcements to the conditional mean and variance equations to isolate the impacts of the target rate news from the effects of the other news.
y t = [RHS of (3.2a) ] + RBANews RBANews t + MacroAnn MacroAnn t

(3.3a) (3.3b)

ln ht = [RHS of (3.2b)] + RBANews RBANews t + MacroAnn MacroAnnt

For the impacts of the RBAs target rate news on the Australian financial markets, in general, it is expected that the sign of the news coefficient in the mean equation to be positive for the debt and exchange rate markets. An unexpected rise in the target rate would have the expected stimulus on the short- and longer-term
39

interest rates. However, the extent to which the shorter- and longer-term rates are affected differently needs to be empirically determined. In addition, as the target rate change represents a change in the real interest rate in the economy, it is expected that an unexpected increase (fall) in the target rate would lead to an appreciation (a depreciation) of the AUD. Thus, a positive sign for the news coefficient for both the interest rate and the exchange rate estimations is expected. The sign of the news coefficient for the stock index returns is expected to be negative as Bernanke and Kuttner (2005) show such a relationship in the US stock market. 3.3.2. The Feds spillover effects on the Australian financial markets For the Feds spillover effects on the Australian financial markets, we introduce the FedNews variable into equation (3.3a) and (3.3b) and lag this variable for one period to account for the timing difference between the US and the Australian markets. Specifically, we estimate equations (3.4a) and (3.4b) as follow.
yt = [ RHS of (3.3a)] + FedNews FedNewst 1
ln ht = [RHS of (3.3b)] + FedNews FedNewst 1

(3.4a) (3.4b)

For the Feds spillover effect, it is expected that unexpected rises in the US Feds target rate would move the Australian interest rates in the same direction, but the stock index and the Australian dollar are expected to fall. As for the conditional volatilities, the response would depend on the informational role the Fed news plays in the Australian markets.

40

3.3.3. The Feds and the ECBs spillover effects on the Asia Pacific financial markets The baseline model presented in equations (3.2a) and (3.2b) has been modified to investigate the Feds and the ECBs spillover effects on the Asia Pacific stock and currency markets in Chapters 5 and 6 accordingly. For the investigations in Chapter 5 on the Asia Pacific stock markets, we keep the Hol dummy to account for the non-synchronous trading across stock markets and drop the Mon dummy. The modified baseline model for our investigations in Chapter 5 is shown in equations (3.5a) and (3.5b) below.
yt = c + Lag yt 1 + Hol Holt + t
ln ht = c + h ln ht 1 + 1

t 1
ht 1

+ 2

t 1
ht 1

+ Hol Holt

(3.5a) (3.5b)

For the empirical tests in Chapter 6 on the Asia Pacific currency markets, we drop the Hol dummy and keep the Mon dummy as currency markets generally do not close except for the weekend. The modified baseline model for our investigation is shown in equations (3.6a) and (3.6b).
yt = c + Lag yt 1 + Mon Mont + t
ln ht = c + h ln ht 1 + 1

t 1
ht 1

+ 2

t 1
ht 1

+ Mon Mont

(3.6a) (3.6b)

To investigate the spillover effects, we include the News variable in both the mean and the variance equations and lag this variable by one period to account for the difference in trading hours between the USs, euro areas markets and the Asia Pacific markets. We also control for the other macroeconomic announcements from the US and the euro area to isolate the target news impacts. However, macro announcement dummies are not included in the conditional mean equations because

41

they represent an average impact of positive and negative macro news (e.g. higher or lower than expected announcements which wound have opposite impacts on stock returns). Macro dummies are stripped of signs, and as such, macro news coefficients would be difficult to interpret. Disaggregated data would be required to unravel the surprise components of each announcement (deviations of actual from expected figures) and investigate the differential impacts of various aspects of announcements on the first moment of market returns.18 However, macroeconomic announcements, irrespective of the types of news contents, could influence market volatility in a similar fashion. For example, unexpectedly higher or lower CPI announcements could both increase the market volatility if they further increase the level of information uncertainty in the market on the days of CPI announcements on average. Alternatively, market volatility could be lower in response to both types of news if the release of new information resolves the heterogeneity in market expectations leading to lower market uncertainty. Thus, the macro announcement dummies are only used to control for the second moment effects on the days of macro announcements. Accordingly, to investigate the Feds and the ECBs news spillover effects on the Asia Pacific stock markets in Chapter 5 we estimate the equations (3.7a) and (3.7b).

18

This is the case when the macro announcement dummies are introduced in both the mean and the

variance equations in empirical tests in Chapter 4. The macroeconomic announcement dummies fail to offer any meaningful explanations in the mean equations.

42

y t = [RHS of (3.5a)] + News News t 1

(3.7a) (3.7b)

ln ht = [RHS of (3.5b)] + News News t 1 + MacroAnn,i MacroAnni ,t 1

i

The target rate news coefficient in the conditional mean equation is expected to be negative for the stock markets as Wongswan (2006, 2009), Bernanke and Kuttner (2005), Ehrmann and Fratzscher (2006), and Rigobon and Sack (2004) documented for various stock markets. Furthermore, unexpected interest rate rises from the Fed and the ECB could have similar market depressing influences in the Asia Pacific region (a negative news coefficient) at least through three factors: (i) higher financing costs for foreign companies that raised capital in the US/euro area markets; (ii) lower stock prices for foreign companies listed in the US/the euro area markets or for those having operations or links with the US/the euro area; and (iii) lower demand from the US/the euro area for foreign commodities (consumption effect). For the spillover effects of the Feds and the ECBs target rate news on the Asia Pacific currency changes, we estimate equations (3.8a) and (3.8b) in Chapter 6. (3.8a) (3.8b)

y t = [RHS of (3.6a)] + News News t 1

ln ht = [RHS of (3.6b)] + News News t 1 + MacroAnn,i MacroAnni ,t 1

i

The sign of the news coefficient in the mean equation can be either negative or positive. In line with empirical findings reported in the literature, the USD and the EUR would appreciate against other currencies in response to unexpected hike in interest rates in these two economies, thus a negative sign is observed.

43

However, as Vickers (2000) finds that the reactions of exchange rates to policy announcements could transmit important insights about the markets perception regarding the central banks credibility in fighting inflation. For instance, an unexpected hike in Feds and the ECBs target rates could be viewed as a signal of future economic contraction and lowering the likelihood of future inflation.19 As a result, such unexpected tightening may lead to a depreciation of the USD and the EUR. In addition, higher interest rates would dampen stock market activities leading to portfolio outflows and this would lead to depreciation of the domestic currency. Therefore, one may observe positive mean coefficients instead. Furthermore, depending on the likely policy response of foreign central banks to the Feds and the ECBs policy decisions, one may even not be able to observe any change in the exchange rates. That is, if market participants believe that the counterpart central banks are more likely to respond to the Feds and the ECBs policy choices, e.g. the counterpart central banks may also unexpectedly increase their own policy interest rates in response to the Feds and the ECBs unexpected actions, the Feds and the ECBs news effect on exchange rates against these correspondent currencies might be muted. In general, the news effect on the conditional volatilities would depend on whether the news adds to or resolves uncertainties in the market. As our volatility specification allows negative coefficients in the conditional volatility equation, it is possible to have both positive and negative sign for the interest rate news coefficient.

19

This is particular true in the case of the ECB under inflation targeting framework.

44

If an unexpected change in the rate leads to further speculation in the market regarding the future direction of the target rate, this heightened heterogeneity would lead to increased trading activities and trading volume, which in turn would lead to higher conditional volatility as empirical evidence shows that trading volume and conditional variance are positively correlated, for example, see Tauchen and Pitts (1983), Joiron (1996), etc. If that is the case, one would observe a positive News. On the other hand, if an unexpected rate change announcement resolves the accumulated heterogeneity in the market and hence reduces the level of uncertainty, a significant volatility reduction would be observed (i.e. a negative News).

45

CHAPTER 4 - THE REACTION OF THE AUSTRALIAN FINANCIAL MARKETS TO THE INTEREST RATE NEWS FROM THE RESERVE BANK OF AUSTRALIA AND THE US FED20

20

A shorter version of this chapter is published in Research in International Business and Finance,

22(3), pp. 378-395, 2008. It has also received the Best paper award at the 2008 International Business and Economics Research Conference, Nevada, USA.

46

4.1. Introduction There is a growing body of literature examining the news contents of central banks interest rate target announcements. The investigations aim at ascertaining the presence and the nature of the news effects on financial markets. Furthermore, as target interest rate announcements are among few events closely watched by market participants, approaching the announcement day, rational investors usually anticipate the direction and the magnitude of such important decision. Accordingly, investors will trade upon their expectation. Therefore, market prices observed immediately prior to the announcement should already incorporate this expectation part. If the announced target rate is different from what already priced, markets react to this surprise (or news) component accordingly. As interest rates of longer maturities are determined by the expected levels of the target rate over the relevant time horizon, any change in the target rate has an immediate impact on other short-term interest rates.21 Thus, central banks influence financial markets through their control over not only the target interest rates but also the markets expectation on the future courses of the respective target rates. The current literature mostly concentrates on investigating the impacts of monetary policy news of the Feds target rate. The Feds interest rate news effects have been investigated on various US markets. For example Bomfim (2003), Bernanke and Blinder (1992), Bernanke and Kuttner (2005) investigate the Feds news effect on the US equity markets. Cook and Hahn (1989), Demiralp and Jorda

21

For example, the 90-day bank accepted bill rates in Australia are calculated as an average of 90

overnights expected interest rates hence.

47

(2004), Gulley and Sultan (2003), Kuttner (2001), Roley and Sellon (1995, 1998), and Lee (2006) examine the Feds news effects on the US debt markets. Recently, attention has also been directed to the spillover effects of the Feds news on foreign financial markets. In particular, the spillover impacts of the Feds interest rate news have been investigated by a number of studies including Bredin et al. (2005) on the Irish stock market, Ehrmann and Fratzscher (2003, 2005a) on the euro area money markets, and Hausman and Wongswan (2006) on the stock, debt and foreign exchange markets. The Feds interest rate news has shown to be transmitted to these markets and the spillover effects are strongly felt. However, the transmission in the opposite direction is found to be weak, see Ehrmann and Fratzscher (2005a). In the Australian context, there exist studies on the Australian market efficiency and the role of news in general on interest rate and exchange markets. For instance, Juttner, Tuckwell and Luedecke (1985) find that the Australian market participants incorporated a comprehensive set of overseas information in setting implied forward rates. Edey and Elliott (1988) find evidence suggested that the Australian Futures markets are efficient where futures prices fully reflect available information. In addition, Tease (1988) finds evidence supporting the expectations theory of the term structure of the Australian short-term interest rates. On the other hand, other studies such as Hogan and Nguyen (1987) and Henry and Summers (1999) focus on exchange rate of the Australian dollar against the US dollar. Turning to the literature on the Australian target interest rate announcement, the RBA started to announce its target interest rate (the overnight cash rate) from January 1998. The RBA Board's decision on the rate is announced in a media release, which states the new target for the cash rate together with the rationale for the

48

decision. However, the empirical literature on the RBAs cash rate announcement effect is limited to the investigation of the announcement impact on the first moment of Australian market returns, e.g. Gasbarro and Monroe (2004) and Diggle and Brooks (2007). Another common limitation of these studies is that they only examine the overall impact of cash rate announcements rather than concentrating on the surprise or news component of the announcement to which markets are responding. Furthermore, the literature is missing a thorough investigation of the spillover effects of the US Feds interest rate announcement news on the Australian financial markets. This is another oversight in the literature as the information leadership role of the US in Australia is well documented. For instance, Kim and Sheen (2000) show that the Australian interest rates (90-day and 10-year rates) react strongly to the first and second moments of the corresponding US rate movements. Masih and Winduss (2006) report a straightforward cointegration relationship between the Australian and the US interest rates, whereas Narayan and Smyth (2004) show similar evidence for the stock markets. Moreover, Kim (2005) reports a direct causal information flow from the US stock market to that of Australia. This chapter has dual aims of firstly investigating the role played by the RBAs interest rate news in the Australian debt, foreign exchange and stock markets, and secondly, documenting and discussing the existence and the nature of the transmission of the US Feds interest rate news on the Australian financial markets. The main findings are summarized as follows. First, there is strong evidence that the RBAs target interest rate news has significant impacts on the daily returns/changes in all three financial market segments in line with prior expectations. In particular, an unexpected rise in the target rate led to a proportional increase in interest rate

49

changes, spot and forward USD/AUD exchange rate returns and two banking stock returns. In addition, the new effects were stronger at the short-term ends of the spectrum for the interest rates and forward exchange rates. This is consistent with the general finding in the literature where the announcement news impact is observed to be weaker at the longer ends. Second, the RBAs target rate news increased the volatility in most cases. This suggests that an unexpected change in the target rate creates further uncertainty regarding future rate changes and hence a higher volatility in most of the markets on the days of target rate announcements. Third, there is weak evidence for an asymmetric news effect where markets tended to have a stronger reaction to unexpected rate rises than unexpected rate cuts. Fourth, the Feds target interest rate news significantly reduced the volatility in the Australian markets following similar volatility reducing effects in the US markets (foreign exchange and stock markets). It can be conjectured that the Feds interest rate news helps reduce the degree of uncertainty in these US markets and this lower volatility environment was transmitted to the Australian markets. The rest of the chapter proceeds as follows. Section 4.2 discusses the nature of the data used in this chapter and the empirical modeling issues are discussed in Section 4.3. Section 4.4 provides the analyses on the empirical investigation results, while Section 4.5 provides robustness check results. Section 4.6 concludes the chapter. 4.2. Data 4.2.1. Target interest rate announcement data The RBA has significantly improved the transparency of its monetary policy since the early 1990s. From January 1990, it started to publicly announce its
50

monetary policy decisions, and from January 1998, the RBA Board's decision to change or leave the cash rate unchanged is announced in a media release at 9:30 am Australian EST (GMT+10) one day following the board meeting.22 The RBA has eleven scheduled Board meetings a year on the first Tuesday of every month except in January. In the case of the US Fed, since 1995, the Feds FOMC publicly announces the Fed funds target rate at 2:00 pm US Eastern Standard Time (GMT-5) unless otherwise specified. The FOMC holds eight regularly scheduled meetings during the year, and other unscheduled meetings as needed. The RBA and the Feds target rate announcement data were obtained from their respective websites.23 Table 4.1 - Descriptive statistics of the RBAs and the Feds target rate news
This table reports descriptive statistics for the RBAs and the Feds target interest rate announcements (Panel A), the news components (Panel B), and the Australian financial market returns/changes series (Panel C) for the period from January 1998 to December 2006. P-values are in braces. The RBAs unexpected component of the cash rate announcements, or news, is calculated as the difference between the actual announcement and the expected component proxied by market consensus reported by the financial press reports sourced from Factiva database. The Feds fund target rate news is calculated using Kuttner (2001) methodology as highlighted in Section 3.2, Chapter 3.

22

The RBA maintains the target interest rate by maintaining the balance in the official money market

at that target rate. It announces its dealing intentions at 9:30 am each day to maintain the balance for the day. From December 2007, a media release is issued at 2.30 pm after each Board meeting. However, this decision does not affect this investigation, which ended at the end of 2006.
23

These data are available at http://www.rba.gov.au/Statistics/cashrate_target.html for the RBA, and

http://www.federalreserve.gov/FOMC/#calendars for the Fed.

51

Panel A. Target interest rate announcements RBA cash rate announcements Total 99 (100%) Rate rise 13 (13%) Rate cut 7 (7%) No change 79 (80%)

Fed fund target rate announcements Total 77 (100%) Rate rise 23 (30%) Rate cut No change 16 38 (21%) (49%)

No. of announcements Proportions

Panel B. Target interest rate news RBA target rate news Total 99 (100%) 0.0135 0.0172 0.0763 3.6837 -0.2500 0.2500 Positive news 6 (6%) 0.2500 0.0000 Negative news 4 (4%) -0.2500 0.0000 No news 89 (90%) Total 77 (100%) -0.0075 0.0031 5.1583 44.4570 -1.9140 0.2760 Fed funds target rate news Positive news 28 (36%) 0.0957 0.0070 0.8946 2.3690 0.0055 0.2760 Negative news 26 (34%) -0.2934 0.1911 -2.2423 8.3325 -1.9140 -0.0050 No news 23 (30%)

No. of observations (Proportion) Summary statistics Mean Variance Skewness Excess Kurtosis Min Max

0.2500 0.2500

-0.2500 -0.2500

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

Panel A of Table 4.1 reports the breakdown of policy announcements into rate rises, rate cuts and unchanged subcomponents. From January 1998 to December 2006, the RBA and the Fed made 99 and 77 target rate announcements, respectively. Of these, the RBA has 20 changed announcements (13 rate rises and 7 cuts) and 79 unchanged announcements. The Fed has 39 changed announcements (23 rises and 16 cuts) and 38 unchanged announcements. Therefore, most of the scheduled interest rate announcements are unchanged decisions (80% for the RBA and 49% for the Fed). 4.2.2. Target interest rate news As the sole instrument for the RBAs monetary policy, the overnight cash rate target affects financial asset prices through conveying new information regarding the RBAs monetary policy stance. The market efficiency hypothesis implies that asset prices observable immediately prior to the RBAs cash rate announcements already incorporate the market expectations on the upcoming announcement. As such, if there is a significant market reaction to the announcement, such an effect must be due to the unexpected (i.e. news) component of the announcement. Thus, in order to gauge the extent to which unexpected changes in a direction or the extent of target

52

rate movements affect financial markets, it is necessary to model properly the news component of the announcements. In order to extract the unexpected component from the RBAs target rate announcements, as briefly discussed in Section 3.2, Chapter 3, financial press reports on the market consensus forecasts has been utilized. Days surrounding the RBAs target rate announcements, the financial press reports what the market consensus was at the time of each announcement. These reports were searched on the Factiva database over a window of a few days before and after the RBAs scheduled announcements. For example, on November 6, 2006 a Factivas report showed that the market expected the RBA would raise 0.25 basis points in its meeting on Wednesday November 8, 2006:
Financial markets will still be hanging on the outcome of the Reserve Bank of Australia's meeting to determine interest rates this week, even though economists say another rate rise is almost a shoo-in. Interest rates are predicted to rise another 0.25 percentage points on Wednesday to 6.25 per cent following the RBA board's meeting tomorrow. It will be the third such rise this year (see Newman, 2006).

The unbiasedness and efficiency of the expectation data generated this way has been tested and the results suggest that the market-based expectations are unbiased predictor of the RBAs interest rate announcements where the null hypothesis cannot be rejected even at 90% level of confidence (see Table 4.2). The news component of each announcement is the difference between the actual target

53

rate change announced and the market expectation.24 The Feds target rate news is generated as detailed in Section 3.2, Chapter 3 using the current-month Fed funds futures contracts in Kuttner (2001) methodology. Table 4.2 - Wald test of the unbiasedness of expectations of monetary policy announcements
This table reports results of the Wald tests for model as described in equation (4.1) for the joint hypothesis Ho: = 0 and =1. P-values are in braces.

ActualChanget = + Expected Changet + t

F-value 0.0230 1.0400 *** 0.4612 {0.5010} {0.0000} {0.7723} Note: *** denotes significance at 1% RBA

(4.1)

Panel B of Table 4.1 reports the summary statistics for the unexpected components of the target rates for both central banks. While 90% of the RBAs announcements contained no news, the US market correctly expected 30% of the Feds announcements. The average unexpected component (or surprise) of the RBAs announcements is 0.0135 percent, whereas the average surprise of the Feds announcements is -0.0075 percent. The variance of the RBAs surprises is also higher than that of the Fed (0.0172 percent for the RBA relative to 0.0031 percent for the Fed).

24

The 90-day bank bill futures rate was also used as the best available proxy for the Australian market

expectations to extract the surprise components as in Kuttner (2001). The estimation results differ slightly from what reported in this chapter using the market expectations and the general conclusion is the same.

54

4.2.3. Australian financial market daily returns Australian interest rate, stock markets and exchange rate data were collected from Datastream for the period from January 1998 to December 2006. Debt markets data consists of short- and long-term Australian interest rates measured as the 90-day bank bill rate, and 3- and 10-year Commonwealth bond yields collected at the Australian market close (4 pm Australian EST (GMT+10)). The exchange rates investigated are mid-level spot, and 1- and 3-month USD/AUD forward exchange rates obtained at the close of London market (5pm GMT).25 For the stock market returns, the overall market index compiled by Datastream is used. In addition, stock prices of the biggest four banks in Australia, namely Australia and New Zealand Banking Corporation (ANZ), Commonwealth Bank of Australia (CBA), National Australia Bank (NAB) and Westpac Banking Corporation (WBC) have been utilized to investigate a disaggregated influence of the target rate surprises on the segment of the stock market that is most directly and immediately affected by the RBAs announcements. 26 Table 4.3 reports the summary statistics of the Australian financial market returns/changes series. The series demonstrate strong evidence of positive skewness for the interest and exchange rate series and negative skewness for the stock series. In all cases, the returns/changes series exhibit leptokurtosis, non-normality and significant serial correlation in the second moment. In addition, all return series

25

The exchange rates are indirectly quoted as one unit of the Australian dollar equals x units of the

US dollar.
26

Daily closing prices are observed at 4pm Australian EST.

55

except for the stock index and the spot exchange rate, significant serial correlation is observed at least at 10%. Table 4.3 - Descriptive statistics of the Australian financial market returns
This table reports descriptive statistics for the Australian financial market returns/changes series for the period from January 1998 to December 2006. P-values are in braces.
90-Day Bank Bill 2370 0.0006 0.0315 0.4665 39.1209 151154 77.7867 {0.0000} 302.2633 {0.0000}
***

No. of observations Mean Std. deviation Skewness Kurtosis Jarque-Bera Ljung-Box Q test Return P-Value Volatility P-Value

Debt market 3 Year Bond 2370 0.0002 0.0665 0.0102 5.2996 2772 42.8172 {0.0022} 230.8850 {0.0000}
***

10 Year Bond 2370 -0.0001 0.0945 0.0268 47.0594 218598 233.2497 {0.0000} 572.0751 {0.0000}
***

USDAUD exchange rate 1-Month 3-Month Spot Forward Forward 2370 2370 2370 -0.0063 -0.0062 -0.0061 0.6888 0.6920 0.7273 0.1591 0.1936 0.0434 2.6521 2.6814 7.0423 704 725 4896 26.9036 {0.1380} 77.8947 {0.0000} 31.9287 {0.0441} 89.4621 {0.0000}
**

***

***

***

***

***

30.9100 {0.0564} 213.7224 {0.0000}

***

No. of observations Mean Std. deviation Skewness Kurtosis Jarque-Bera Ljung-Box Q test Return P-Value Volatility P-Value

ANZ 2370 0.0450 1.2727 -0.1054 1.3716 190 31.1432 {0.0533} 668.9714 {0.0000}
*

CBA 2370 0.0441 1.1165 -0.2545 1.9232 391 32.1658 {0.0416} 228.9215 {0.0000}
**

Stock market NAB 2370 0.0303 1.2582 -0.8623 8.5894 7576 50.1271 {0.0002} 251.4905 {0.0000}
***

WPC 2370 0.0399 1.2013 -0.1127 1.1252 130 32.1544 {0.0417} 470.6277 {0.0000}
**

Stock Index 2370 0.0355 0.7562 -0.3006 3.6918 1381 12.8212 {0.8849} 239.8054 {0.0000}

***

***

***

***

***

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

4.3. Empirical modeling issues 4.3.1. Baseline EGARCH(1,1) model The econometric literature shows that the GARCH family of models is well suited to modeling daily financial returns series, which are characterized as skewed, leptokurtic and non-normal distributions with time-varying second moments as shown in Panel C, Table 4.1 for those variables used in this investigation. This section starts with the baseline univariate EGARCH(1,1) model and then progress to specific modeling of the interest rate news effects of various forms of the RBAs and the Feds news on the Australian financial markets.

56

The baseline EGARCH(1,1) model employed in this study can be described by the conditional mean and the conditional variance equations (4.2a) and (4.2b) shown below.
p

y t = c + Lag ,i y t i + Mon Mon t + Hol Hol t + t

i =1

(4.2a) (4.2b)

ln ht = c + h ln ht 1 + 1

t 1
ht 1

+ 2

t 1
ht 1

+ Mon Mont + Hol Hol t

The conditional mean equation for the returns/changes in the financial market series ( yt ) is expressed as a function of past returns as well as Monday effect (Mon), Holiday effect (Hol) in relevant markets. The conditional variance equation for the returns/changes in the financial market series ( ht ) is expressed as a function of the past variance, Monday effect (Mon), and holiday effect (Hol). The Monday effect variable takes the value of one for Mondays and zero otherwise. The Holiday effect variable is specific to each market segment and records the number of days of market closure between two successive market prices. For example, for normal consecutive daily observations (i.e. returns calculated over two days - Monday close to Tuesday close price) the value of zero is assigned, whereas values of one or higher value will be assigned for returns observations calculated over a longer horizon due to market closure. These seasonal dummies help to control for the days of more intense information flows following a longer period of closure. 4.3.2. Overall news effects of the RBAs target interest rate news The base model shown in (4.2a) and (4.2b) is then augmented with various forms of interest rate news variables. The first model includes an overall RBA news variable, RBANewst, as described in equations (4.3a) and (4.3b) below to investigate

57

the overall news effects of the RBAs target news on the first and second moments of returns. In addition, there was at least one scheduled macroeconomic announcement made by the Australian Bureau of Statistics (ABS) on the same day as the RBAs target rate on 42 out of the 99 days.27 The RBAs announcements were made at 9.30 am while the ABSs announcements were at 11.30 am Australian EST. Therefore, it is necessary to control for these announcements to isolate the RBAs interest rate news effect on the daily returns of the Australian financial markets on these days. A dummy variable, MacroAnnt is added to both the conditional mean and the variance equations that take the value of one for those RBAs announcements that coincided with the release of other macroeconomic variables on the same day, and zero otherwise.
y t = [RHS of (4.2a)] + RBANews RBANews t + MacroAnn MacroAnn t

ln ht = [RHS of (4.2b)] + RBANews RBANewst + MacroAnn MacroAnnt

(4.3a) (4.3b)

In general, it is expected that the sign of the news coefficient in the mean equation to be positive for the debt and the foreign exchange markets. An unexpected rise in the target rate would have the expected stimulus on the short- and longer-term interest rates. However, the extent to which the shorter- and longer-term rates are affected differently needs to be empirically determined. As the 90-day rate is based on the target cash rate, it is expected that the target cash rate changes would result in

27

Four different types of macroeconomic variables were announced in a total of forty-two days when

the RBAs cash rate announcements were made. These are CPI inflation, employment, international accounts and retail sales. Details of the release dates are shown in Appendix A.

58

a same direction move in the 90-day rate. For the longer-term rates, a change in the same direction as the news effect is expected to be transmitted from the short ends to the long ends. In addition, as the target rate change represents a change in the real interest rate in the economy, it is expected that an unexpected increase (fall) in the target rate would lead to an appreciation (a depreciation) of the AUD. Thus, a positive sign for the news coefficient for both the interest rate and the exchange rate estimations is expected. The sign of the news coefficient for the stock index returns is expected to be negative, in general, as Bernanke and Kuttner (2005) show such a relationship in the US stock market. For the banking stocks, depending on their balance sheet exposure to interest rate changes, either positive or negative coefficient might be observed. The news effect on the volatilities would depend on whether the news adds to or resolves uncertainties in the market. If an unexpected change in the rate leads to further speculation in the market regarding the future direction of the target rate, this increased heterogeneity would be shown as a positive news coefficient in the variance equation, i.e. a rise in the volatility. On the other hand, a market calming effect could be observed if an unexpected change resolves uncertainty and this is shown as a negative news coefficient. Table 4.4 summarizes a priori expectations for the RBAs target news effects on the Australian financial markets. Table 4.4 - A priori expectations
This table shows expected signs for the coefficients on the conditional mean and volatility equations of (4.3a) and (4.3b). +/- denote positive/negative coefficients Debt markets USD/AUD exchange rates Stock markets RBA's effects Mean Volatility + +/+ +/+/-

59

4.3.3. Asymmetric effects of the RBAs target rate news There is a potential for asymmetric impacts of unexpected interest rate rises and cuts. The announcement news literature reports bad news usually have larger impact. For example, Connolly and Wang (2003) report that volatility spillovers amongst the US, the UK and Japan depend on whether the announcement was good or bad news.28 In particular, bad news from the UK and the US, e.g. higher than expected unemployment rate, lead to significant increases in volatility in Japan. For monetary policy announcements, Bomfim (2003) finds that unanticipated rises in the Fed Funds target rate tend to have a larger effect on the US stock market volatility than unanticipated cuts. Furthermore, Gulley and Sultan (2003) find that US monetary policy news had an asymmetric effect on the stock market, but a symmetric effect on the bond market. The asymmetric influence of unexpected increases and unexpected cuts of the RBAs target rate news is tested by disaggregating the overall news variable into two types of news, unexpected rises and unexpected cuts. The RBANewst variable is partitioned into a rate rise surprise component, RBANews_Rt, and a rate fall,
RBANews_Ft, and added to equations (4.2a) and (4.2b) as shown in equations (4.4a)

and (4.4b) below. As suggested in the literature, it is expected that the Australian

28

The announcements examined are industrial production, unemployment rate, merchandise trade

balance, money supply, consumer price inflation, and producer price inflation for the US; M3 (19851990) and M4 (1991-1996) money supply, retail price inflation, industrial production, and the unemployment rate for the UK; and money supply, industrial production, consumer price index, and wholesale price index for Japan.

60

market would react more strongly to unexpected interest rate increases than cuts. The potential for an asymmetric effect is ascertained by examining the difference between the two coefficients in terms of signs and magnitudes.
y t = [ RHS of (4.2a)] + RBANews _ R RBANews _ Rt + RBANews _ F RBANews _ Ft +

(4.4a)

MacroAnn

MacroAnnt

ln ht = [RHS of (4.2b)] + RBANews _ R RBANews _ Rt + RBANews _ F RBANews _ Ft +

(4.4b)

MacroAnn MacroAnn t

4.3.4. Spillover effects of the US Feds target interest rate news The base model is then modified to examine the spillover effects of the US Feds target rate news on the Australian financial markets. Both direct and indirect transmission channels of the US news are reported in the literature. To the extent that the US stock markets respond to the Feds interest rate news as documented in the current literature, and the US markets lead other markets such as Australia, e.g. studies of Narayan and Smyth (2004) and Kim (2005) for stock market linkages, and similar evidence is shown in the Australian interest rate movements as in Kim and Sheen (2000), the Feds interest rate news can have an indirect influence on foreign markets. For example, the Australian stock markets would take a lead from the overnight US market movements that already incorporate the US markets responses to the Feds news. A first step trying to investigate the direct influence of the US Feds interest rate announcements on other countries is shown in Hausman and Wongswan (2006). This chapter aims to provide more detailed analyses on the direct channel of influence by investigating the extent of the Australian markets responses to the Feds news announcements.

61

This spillover effect is explicitly modeled in both the conditional mean and volatility of the daily Australian market returns/changes. An overall Feds surprise variable (FedNews) is introduced into models (4.3a) and (4.3b). The FedNews variable is lagged by one period to account for the time difference between the US and Australia. The overall Feds news effects are detected via the following system of conditional mean and variance equations.
y t = [ RHS of (4.3a)] + FedNews FedNewst 1
ln ht = [RHS of (4.3b)] + FedNews FedNews t 1

(4.5a) (4.5b)

It is expected that the unexpected rise in the US Feds target rate would move the Australian interest rates in the same direction, but the stock index and the Australian dollar are expected to fall. As for the conditional volatilities, the response would depend on the informational role the Fed news plays in the Australian markets. In addition, the literature reports that worse than expected monetary policy news tended to have bigger impacts in the US markets, see e.g. Bomfim (2003), Gulley and Sultan (2003) and Connolly and Wang (2003). This, along with the findings of the US spillover impact on the Australian markets, requires an investigation of the potential asymmetric spillover impact of the Feds news on the Australian markets. The FedNews variable is disaggregated into two subcomponents: unexpected rise (FedNews_R) and unexpected fall (FedNews_F). The

62

following system presents the conditional mean and conditional variance equations used to detect such asymmetric spillover effects.29
y t = [RHS of (4.4a)] + FedNews _ R FedNews _ Rt 1 + FedSur _ F FedNews _ Ft 1

(4.6a) (4.6b)

ln ht = [RHS of (4.4b)] + FedNews _ R FedNews _ Rt 1 + FedSur _ F FedNews _ Ft 1

4.4. Empirical results 4.4.1. The baseline EGARCH(1,1) results The quasi-maximum likelihood estimations of the baseline EGARCH(1,1) model described by equations (4.2a) and (4.2b) are reported in Table 4.5. Two lags of the dependant variable are employed and this addresses the residual serial correlations in most cases.30 In the conditional mean equation for the interest rate changes, the coefficients of the first and the second lags are significant and negative. However, only the first lag coefficient is significant and it is positive for the stock returns of CBA, NAB and WBC. The Mon dummy is significantly negative for the 10-year rate changes, spot and 1-month forward USD/AUD exchange rate returns. The Hol dummy is positive and significant for the CBA stock return and 3- and 10-year bond rate changes.

29

As a robustness check, two components of the Fed news variables were added to (4.2a) and (4.2b)

instead, however the resulting Fed surprise coefficients are essentially the same as what reported here.
30

For the remaining serial correlations in the first and second moments of the standardized residuals,

varying the lags of the dependant variables and EGARCH lags addressed the remaining correlations. However, this did not have material impact on the estimation results of the interest rate news variables. Thus, for consistency across models, the same specification is used throughout.

63

In the conditional variance equations, the lagged variance term (h) is close to one in all cases except for the 10-year rate where it is negative, suggesting volatility persistence as found in the literature. Evidence for the volume effects of innovations has been found, that is, unexpected changes, regardless of the direction, in the mean has a significant impact of raising the conditional volatility (positive 2). There is also some evidence for an asymmetric effect of the innovations. A negative influence (1<0) is found for the NAB, stock index returns, 3-year bond rate changes, and 3month forward exchange rate returns. This suggests that an unexpected fall in the conditional mean in these cases leads to even higher conditional variances. On the other hand, a positive impact (1>0) is shown for the 90-day and the 10-year interest rate changes. The conditional variances are lower, in general, on Mondays, while they are generally higher on the days immediately following market closure due to holidays.

64

Table 4.5 - EGARCH(1,1) estimations of daily returns

This table reports the quasi-maximum likelihood estimates of the EGARCH (1,1) model as described in equations (4.2a) and (4.2b) of daily Australian interest rate changes, USD/AUD exchange rate returns and stock market returns. P-values are in braces.
p

y t = c + Lag ,i yt i + Mon Mont + Hol Holt + t

i =1

(4.2a) (4.2b)

ln ht = c + h ln ht 1 + 1 ht 1 ht 1

t 1
+ 2 + Mon Mont + Hol Hol t

t 1

Panel A Mean equation

10 year 0.0376 {0.0552} 0.0490 {0.1219} -0.0075 {0.7264} -0.0036 {0.9309} 0.0081 {0.9115}
*

90-day

Debt markets 3 Year ANZ CBA 0.0736 *** {0.0000} 0.1150 *** {0.0000} -0.0139 {0.4840} 0.0432 {0.2459} 0.1087 * {0.0773}

Stock Market NAB 0.0388 * {0.0823} 0.0890 *** {0.0000} 0.0039 {0.8536} 0.0334 {0.4294} -0.0860 {0.2356}

WPC 0.0547 ** {0.0437} 0.0756 *** {0.0002} -0.0040 {0.8448} -0.0285 {0.5032} 0.0919 {0.2773}

Index 0.0208 *** {0.0051} 0.0105 {0.5827} -0.0052 {0.7811} 0.0323 {0.3228} 0.0840 {0.2048}

USDAUD exchange rates Spot 1 m forward 3 m forward Conditional mean equation 0.0006 *** 0.0020 *** 0.0013 0.0201 0.0212 0.0150 c {0.0048} {0.0096} {0.1211} {0.1106} {0.1042} {0.3277} -0.0593 *** -0.0406 ** -0.0445 ** 0.0031 0.0048 -0.0179 Lag, 1 {0.0000} {0.0348} {0.0158} {0.8821} {0.8166} {0.2775} 0.0516 *** -0.0422 ** -0.0387 *** -0.0011 -0.0068 -0.0050 Lag, 2 {0.0018} {0.0355} {0.0027} {0.9584} {0.7458} {0.8066} 0.0001 -0.0029 -0.0116 *** -0.0525 * -0.0545 * -0.0594 Mon {0.8073} {0.2593} {0.0014} {0.0752} {0.0715} {0.1220} -0.0001 0.0146 ** 0.0199 ** -0.0205 -0.0341 -0.0450 Hol {0.9830} {0.0407} {0.0235} {0.7841} {0.6684} {0.5474} Note: *, **, *** denote significance at 10%, 5%, and 1%, respectively

65

Table 4.5 Continued

Panel B Variance equation

90-day
*** *** *** ***

Debt markets 3 Year 10 year ANZ CBA

Stock Market NAB

WPC
***

Index
***

c
*** *** ***

***

***

h
***

***

***

***

***

***

1
*** ***

***

***

***

2
*** ***

***

***

***

***

***

***

Mon
***

***

***

***

***

***

Hol

-0.3846 {0.0000} 0.9720 {0.0000} 0.0494 {0.0000} 0.2973 {0.0000} -0.1522 {0.0000} 0.6296 {0.0000}

***

-0.2318 {0.0000} 0.9781 {0.0000} -0.0189 {0.0000} 0.1477 {0.0000} -0.0060 {0.3533} 0.2153 {0.0000}

***

-6.7607 {0.0000} -0.2168 {0.0000} 0.0550 {0.0000} 0.6303 {0.0000} 0.4962 {0.0000} 0.5565 {0.0000}

-0.0976 {0.0000} 0.9698 {0.0000} -0.0050 {0.6755} 0.1758 {0.0000} -0.1842 {0.0000} 0.0171 {0.6291}

-0.0996 {0.0000} 0.9727 {0.0000} 0.0095 {0.2944} 0.1630 {0.0000} -0.1502 {0.0000} -0.0357 {0.4047}

-0.0949 {0.0000} 0.9772 {0.0000} -0.0168 {0.0527} 0.1538 {0.0000} -0.1404 {0.0000} 0.0791 {0.0104}

**

-0.1171 {0.0000} 0.9650 {0.0000} 0.0157 {0.1430} 0.1834 {0.0000} -0.1426 {0.0000} 0.1056 {0.0087}

***

-0.1528 {0.0000} 0.9804 {0.0000} -0.0918 {0.0000} 0.1029 {0.0000} 0.2813 {0.0000} 0.0916 {0.0077}

***

LogL Q(20)
** ***

5575 34.6662 {0.0220}

**

3178 27.7139 {0.1163}

2696 33.5644 {0.0292}

USDAUD exchange rates Spot 1 m forward 3 m forward Conditional variance equation -0.0368 *** -0.0222 *** -0.0859 *** {0.0000} {0.0000} {0.0000} *** *** 0.9937 0.9971 0.9933 *** {0.0000} {0.0000} {0.0000} -0.0040 0.0024 -0.0053 ** {0.3810} {0.5653} {0.0210} *** *** 0.0633 0.0586 0.0819 *** {0.0000} {0.0000} {0.0000} -0.1012 *** -0.1414 *** 0.0896 *** {0.0000} {0.0000} {0.0000} 0.2269 *** 0.2223 *** 0.1690 *** {0.0000} {0.0000} {0.0000} Estimation diagnostics -2401 -2406 -2505 24.4674 29.4178 * 28.6472 * {0.2226} {0.0799} {0.0949} -3434 16.3243 {0.6963} -3196 23.5822 {0.2611} -3272 19.9796 {0.4592}

-3353 24.0919 {0.2384}

-2493 11.1004 {0.9436}

18.0505 90.7487 5.4240 18.4343 23.8941 87.3449 16.6027 16.9941 27.2085 14.6040 18.0531 {0.5841} {0.0000} {0.9995} {0.5588} {0.2470} {0.0000} {0.6786} {0.6534} {0.1295} {0.7986} {0.5839} Note: *, **, *** denote significance at 10%, 5%, and 1%, respectively. Q(20) and Q2(20) are Ljung-Box portmanteau test of white noise for linear and non-linear (squared) standardized residual.

Q2(20)

***

66

4.4.2. Overall effect of the RBAs target interest rate news The quasi-maximum likelihood estimates of the EGARCH model for the overall effect of the RBAs target rate news as modeled in equations (4.3a) and (4.3b) are reported in Table 4.6. Significant news effects on interest rates and the USD/AUD exchange rates are detected. A significant and positive news coefficient is found in the conditional mean equations of the 90-day and the 3-year rate estimations. In response to a onepercentage point (25 basis points) unexpected rise in the target rate, the 90-day bank bill rate rose by 0.3601 percent (0.0900 percent) while the 3-year bond rate was increased by 0.1570 percent (0.0393 percent). The AUD appreciated when the RBA announced an unexpected rate rise. In response to a one-percentage point (25 basis points) unexpected rise in the target rate, the spot rate appreciated by 2.7074 percent (0.6769 percent) while the 1- and 3-month forward rates appreciated by 1.8744 (0.4686 percent) and 1.7157 percent (0.4289 percent), respectively. Remarkably, the impact of the RBAs news on the first moment of interest rate and exchange rate changes are larger in magnitude at the short-term spectrum. This finding is consistent with previous studies, such as Cook and Hahn (1989)s finding that an increase of 50 to 55 basis points in the short-term T-bill rates in response to a one basis point increase in the Fed funds target rate, but only a 10 basis point increase in the 20-year bond yield is observed.

67

The interest rate news also has a significant impact on the conditional volatilities in nearly all cases. Unexpected announcements led to a higher volatility on the announcement days in most cases. This result is similar to Ehrmann and Fratzscher (2003) where the Bundesbanks interest rate announcement increased the volatility of the German money markets prior to 1999. This might be because in approaching an impending RBAs target rate announcement, market participants might correctly guess the direction of the possible target rate change but are unsure about the magnitude of such an action. Therefore, when there is a surprise in the RBAs announcement the homogeneity evaporates and instead would lead to heterogeneity of beliefs among market participants. This then leads to a higher volatility in most markets on the announcement day. Another possible explanation might be on the quality and effectiveness of the RBAs policy communications towards improving the transparency and accountability of its monetary policy. For the stock market series, the stock prices of NAB and WBC significantly rose in response to the news. In response to a one-percentage point (25 basis points) unexpected rise in the target rate, the NABs stock rose by 1.9621 percent (0.4905 percent) and the WBCs by 1.9372 percent (0.4843 percent). This can be conjectured that this might be due to both NAB and WBC having more interest rate sensitive assets than liabilities denominated in the same currency for most of the period under investigation. As a result, an unexpected rise in the target rate would be expected to

68

have a positive influence on these banks incomes and this would be shown as a current rise in their stock prices.31 There is also evidence for a significant volatility effect in some cases. The interest rate news raised the conditional volatility of the returns of ANZ, WBC and the stock index. The macroeconomic announcement dummy is shown to be ineffective in explaining the mean equations. 32 However, it is contributing to the conditional variance equations. In three out of the five cases where a significant coefficient is detected, the macroeconomic announcements significantly reduced the conditional volatility.

31

For example, the net interest rate sensitive asset positions of the NAB at balance sheet dates for the

period 2002 to 2006 were such that it would have experienced a net income rise by 21, 67, 39, 50 and 42 million AUD, respectively, in response to a one percent parallel rise in interest rates across the whole Australian yield curve. For WBC, for the same period, it would have experienced an income rise of 17, 29, 30 and 26 billion for the period 2002-2005, respectively, and income deterioration by 32 billion AUD in 2006. Although ANZ also had a similar risk position for a number of years, the RBAs news did not have a noticeable impact.
32

This is not a surprise result as the macro news dummies are stripped of signs and thus represent an

average impact of positive and negative news (e.g. higher or lower than expected announcements which wound have opposite impacts on stock returns).

69

Table 4.6 - The RBAs news effects on the Australian financial markets

This table reports the quasi-maximum likelihood estimates of the EGARCH model as described in equations (4.3a) and (4.3b) of daily Australian stock market returns, interest rate changes, and USDAUD exchange rates. P-values are in braces.

y t = [RHS of (4.2a)] + RBANews RBANews t + MacroAnn MacroAnn t

(4.3a) (4.3b)

ln ht = [RHS of (4.2b)] + RBANews RBANewst + MacroAnnMacroAnnt

90-day 1.5636 {0.2434} -0.0026 {0.9845}

***

Debt markets 3-Year ANZ 0.9286 {0.3891} -0.0374 {0.7841} CBA

10-year

Spot

Stock Market NAB WPC 1.9621 {0.0057} -0.1421 {0.2814}

***

Index 1.9372 {0.0207} -0.0590 {0.6332}

**

-0.6784 {0.5860} -0.0818 {0.3433}

*** ***

USDAUD exchange rates 1M Forward 3M Forward Conditional mean equations RBASur 0.3601 *** 0.1570 * 0.0124 2.7074 ** 1.8744 * 1.7157 ** {0.0000} {0.0571} {0.9161} {0.0414} {0.0658} {0.0124} MacroAnn -0.0034 0.0003 0.0126 0.0238 0.0175 0.0652 {0.0520} {0.9735} {0.0111} {0.8054} {0.8694} {0.5099} Conditional variance equations RBASur 1.2396 *** 2.5810 *** 0.7065 1.1078 *** 0.9516 *** -2.4968 {0.0049} {0.0000} {0.7199} {0.0000} {0.0000} {0.1943} MacroAnn -0.2143 *** 0.2016 *** -0.5544 *** -0.0656 ** -0.0020 -0.1440 {0.0000} {0.0000} {0.0006} {0.0441} {0.9037} {0.5657} Note: *, **, *** denote significance at 10%, 5%, and 1%, respectively 1.4728 {0.0015} -0.1042 {0.1296} 0.5816 {0.4549} 0.0086 {0.9121}

0.8685 {0.1368} -0.0413 {0.5974}

1.6610 {0.0017} -0.0499 {0.4315}

1.4323 {0.0000} 0.3432 {0.0000}

***

70

4.4.3. Asymmetric impacts of the RBAs target interest rate news Table 4.7 shows the disaggregated news coefficients of the conditional mean and variance equations of (4.4a) and (4.4b). For the debt market, the two long-term interest rates responded only to unexpected rises whereas the 90-day rate responded to both types of news. In all cases, interest rates rose significantly to the news. In the case of the exchange rates, the spot rate does not show an asymmetric response, however, the 1-month forward rate responded only to unexpected cuts whereas the 3month forward responded only to unexpected rises. There is no asymmetric response shown in the stock market. The evidence on the asymmetric impacts on the conditional volatility is less uniform. In general, only unexpected target rate cuts raised the interest rate volatilities, whereas the reverse is true for the exchange rates. For the stock market, all the significant coefficients have a positive sign, suggesting a volatility increase. Unexpected rises led to a higher volatility, in general, although unexpected cuts also increased the volatility for the CBA and the stock index returns. Furthermore, in most cases, the hypothesis that the magnitudes of the two different types of news coefficients are equal in the variance equation is rejected. Thus, there is a complex array of volatility responses to each type of news.

71

Table 4.7 - The RBAs asymmetric surprises effects on the Australian financial markets

This table reports the quasi-maximum likelihood estimates of the EGARCH model as described in equations (4.4a) and (4.4b) of daily Australian stock market returns, interest rate changes, and USDAUD exchange rates. P-values are in braces. (4.4a) y t = [RHS of (4.2a)] + RBANews _ R RBANews _ Rt + RBANews _ F RBANews _ Ft + MacroAnn MacroAnnt (4.4b)

ln ht = [RHS of (4.2b)] + RBANews _ R RBANews _ Rt + RBANews _ F RBANews _ Ft + MacroAnn MacroAnn t

10-year 2.2871 {0.5120} -0.1443 {0.9450} -0.1822 {0.1566} Spot Stock Market NAB WPC

90-day

Debt markets 3-Year

Index 0.6335 {0.6129} 2.6617 {0.3897} -0.0630 {0.6053}

***

-1.0126 {0.4616} 2.1456 {0.3847} -0.0816 {0.3594}

***

0.8330 {0.3092} *** 0.8072 {0.3884} -0.0416 {0.5294}

2.5964 {0.0000} -0.0614 {0.9705} -0.0767 {0.2047}

0.0932 {0.9072} 2.5837 {0.0000} -0.2995 {0.0000}

***

USDAUD exchange rates 1M Forward 3M Forward ANZ CBA Conditional mean equations RBASur_P 0.3883 *** 0.2807 ** 0.1360 ** 3.5535 3.1317 2.7929 *** -0.5100 0.0796 {0.0000} {0.0107} {0.0219} {0.1404} {0.1202} {0.0000} {0.7281} {0.9635} RBASur_N 0.3386 *** -0.0831 -0.1602 0.2998 0.2693 *** 0.2179 3.6330 3.1339 {0.0000} {0.6210} {0.4973} {0.9031} {0.0019} {0.8251} {0.3084} {0.4256} MacroAnn -0.0029 * -0.0037 -0.0030 0.0106 0.0103 0.0294 0.0066 -0.0047 {0.0955} {0.7155} {0.7416} {0.9125} {0.9023} {0.7539} {0.9615} {0.9759} Conditional variance equations RBASur_P -0.5235 -0.2360 -4.6372 ** 1.5304 *** 1.5960 *** -4.5705 1.5898 *** -0.9814 {0.4646} {0.6119} {0.0209} {0.0000} {0.0000} {0.2405} {0.0097} {0.3423} RBASur_N 4.0801 *** 5.8815 *** 2.5300 0.3235 -0.0445 -3.2599 0.9574 2.4105 {0.0000} {0.0000} {0.3545} {0.4311} {0.8925} {0.3107} {0.4662} {0.0016} MacroAnn -0.1720 *** 0.1527 *** -0.5665 *** -0.0819 ** -0.0281 ** -0.1737 -0.1217 ** 0.0487 {0.0011} {0.0000} {0.0000} {0.0103} {0.0296} {0.4674} {0.0433} {0.5395} Tests of equality between positive and negative news coefficients 0.5747 3.2937 * 1.3912 0.8772 1.9316 5.1548 ** 1.1532 0.4991 Mean Eq {0.4484} {0.0695} {0.2382} {0.3490} {0.1646} {0.0232} {0.2829} {0.4799} 24.3623 *** 104.3811 *** 4.7967 ** 6.4158 ** 17.0581 *** 0.0623 0.1764 7.0601 Var Eq {0.0000} {0.0000} {0.0285} {0.0113} {0.0000} {0.8028} {0.6745} {0.0079} Note: *, **, *** denote significance at 10%, 5%, and 1%, respectively 0.3870 {0.5339} *** 0.0005 {0.9831}

0.3414 {0.5590} 2.2775 {0.1313}

1.2494 {0.2637} 7.6813 {0.0056}

***

72

4.4.4. Spillover effects of the US Feds target interest rate news This section shows evidence for the spillover the impact of unexpected Fed funds target rate movements on the Australian financial markets.33 The results are reported in Table 4.8. 34 There is no evidence of the spillover effect on the daily Australian interest rate changes, however, the USD/AUD exchange rate responded as per prior expectations. Both spot and one-month forward rates depreciated (appreciated) in response to an unexpected US Fed target rate rise (fall). In addition, the Australian stock index significantly fell following the US interest rate news. This might be because an interest rate rise in the US would eventually dampen the US and the world demand for commodities and resources, and hence a negative influence on the Australian stock markets where resource stocks are influential. More importantly, it has been found that the Feds interest rate news injected market-calming influence in all three segments of the Australian markets, while the RBAs news resulted in higher levels of volatility as reported above. This is contrary

33

Weak evidence (at 10%) has been found that the US Fed target rate changes Granger cause the

RBAs cash rate changes with two lags included in the test equations, but the causality does not run in the other direction. See Appendix B.
34

Investigations to ascertain the nature of the US Feds target interest rate news on the US markets

has also been conducted. The short-term interest rate (90-day T-Bill) moved in the same direction as the unexpected target rate movement and the volatilities of the USD exchange rates and stock indices were significantly reduced. Details of the estimations are shown in Appendix C.

73

to what Bredin et al. (2005)s finding where the US Feds announcements generally increased the Irish equity market volatility on the days of announcement. In particular, they report that an unexpected increase of the Feds target rate increased the volatility. There are at least two possible explanations for the market calming effects of the US Feds news on the Australian markets. First, as shown in Appendix C, the US Feds target interest rate announcement news significantly reduced the daily return volatilities of spot and forward USD exchange rates and stock indices. Given the established pattern of information leadership of the US, the momentum of the trading environment in the US following the Feds news announcement was transferred to the Australian markets leading to lower market volatilities. Second, the Australian markets might be at an information disadvantage regarding forming expectations on the US Feds interest rate decisions compared to the US market participants. Thus, a relatively larger proportion of under-informed traders in Australia would have caused a higher diversity of opinions on the impending US announcements compared to the RBAs announcements where they would have a relatively clearer position. However, since the Australian markets open for trading after the US markets close on a calendar day, Australian market participants observe any heterogeneity of opinions dissolved over the course of the US trading day after the Feds interest rate news announcements. By the time the Australian markets open, a clear US market movement would have been established and the implications of the Feds news on the Australian markets would have been well understood. Thus, the Feds news would have injected market-calming influences leading to significant reductions in the conditional volatilities of market returns/changes in the Australian financial markets.

74

4.4.5. Asymmetric spillover effects of the US Feds target interest rate news In general, no strong evidence for asymmetric influences of the Feds news on the Australian markets has been detected. Nonetheless, it is observed that, the USD/AUD exchange rates tend to depreciate more in response to unexpected US rate cuts, and the volatility reducing effect is bigger with unexpected US rate rises on the Australian interest rates (see Table 4.9). 4.5. No-news effects As the measure of monetary policy surprise for the RBA is based on the market consensus and thus might mask possible heterogeneity in the expectations. A monetary policy decision that is fully expected by the mean or median market participant can nonetheless resolve or create market uncertainty and thus the confirmation of the markets beliefs could still have a material impact on the conditional volatility of market returns. A Nonews dummy that takes the value of one for those RBA announcements that were fully expected by the market (i.e. there is no news) and zero otherwise is added to the conditional volatilities in the models reported in Tables 4.6 to 4.9. Panel A of Table 4.10 show estimation results for the RBAs target rate news effects with Nonews dummy in the conditional variance equation. Although this variable does not have any significant impact on the estimations reported in Table 4.6, significantly higher conditional volatilities of all three interest rates, the NABs stock returns and the stock index returns on these days have been clearly shown. However, opposite result for the exchange rates, in general, has been found. In addition, as shown in Panel B of Table 4.10, while the results reported for the RBAs asymmetric impacts as in Table 4.7 are generally unchanged, the Nonews dummy shows significant higher volatility in debt markets and three

75

stock markets including the ANZ, NAB and the general stock market index. On the other hand, lower volatility has been shown in 2 out of 3 exchange rate returns. Similar to the RBAs news effects, as can be seen from Panels C and D of Table 4.10, the inclusion of the Nonews dummy does not significantly affect the Feds target news spillover effects reported in Tables 4.8 and 4.9. The Nonews dummy shows significantly higher volatility has been detected for all debt markets, some stock market returns while lower volatility has been found for most exchange rate changes, in general. 4.6. Conclusion This chapter provides comprehensive evidence on the impacts of the RBAs and the US Feds target interest rate news on various segments of the Australian financial markets. A number of important findings have been found. First, the RBAs target interest rate news has statistically significant impacts on the conditional means of the daily returns/changes in the Australian debt, foreign exchange and stock markets. Unexpected rate rise announcements significantly raised the 90-day bank bill and the 3-year government bond interest rates, appreciated the USD/AUD exchange rate in the spot, 1- and 3-month forward markets, and stimulated stock returns of two banks, NAB and WBC. In addition, the RBAs news effect is stronger at the short-term ends of the interest rates and forward exchange rates. Second, the RBAs news raised the level of volatility in most cases. Apparently, when markets were caught by surprise the implications of the changes were interpreted differently in the markets. The resulting heterogeneity of opinions would have led to an increased volume of trade and hence higher volatilities on the days of the announcements with an unexpected target rate change. Third, some evidence for

76

asymmetric effects of policy surprises has been shown. Markets tend to respond more strongly to unexpected rate rises than rate cuts, in general. Fourth, the US Feds target rate news significantly reduced the volatility in the Australian markets. This might be due to at least two reasons, first the market momentum in the US of the volatility reducing effect of the Feds news was spilt over to the Australian market leading to similar reductions in volatility, and the Feds news resolving the level of uncertainty in Australia regarding the US economy which has a direct bearing on that of the Australias. As the RBAs monetary policy aims at anchoring the market expectation and maintain the inflation target of 2-3 percent per annum, these findings help the RBAs decision makers to have a better understanding of potential impacts of the RBAs and the Feds monetary policy news on the Australian financial markets, especially the finding that while the RBAs news raised volatility in the Australian financial markets, the Feds news significantly reduced the volatility level would help the RBA to assess its monetary policy effectiveness. This, in turn, helps the RBA to steer the economy in the desired target by improving the monetary policy conduct and adjusting the policy stance according to the market responses. This chapters findings are also useful to market participants in designing trading mechanisms as well as risk management strategies in response to both domestic and external interest rate shocks. For example, understanding that an unexpected increase in the RBAs target interest rate would raise the 3-year Commonwealth bond interest rates, market participants would make profits if they could short-sell the 3-year bond before the RBAs meeting day and buy back at a lower price as a result of a higher interest rate. In addition, interest rate risk is perceived as the biggest risk that most of financial institutions faces due to the

77

inherent mismatches in the assets and liabilities structure, forward looking hedging strategies building on the understanding of potential impacts of the central banks target rate would help financial institutions not only to reduce the likelihood of making loss but also increase the chance of making profits.

78

Table 4.8 - Spillover effects of the Feds target rate news in the Australian financial markets

This table reports the quasi-maximum likelihood estimates of the EGARCH model as described in equations (4.5a) and (4.5b) of daily Australian stock market returns, interest rate changes, and USDAUD exchange rates. P-values are in braces.
yt = [RHS of (4.3a)] + FedNews FedNewst 1

ln ht = [RHS of (4.3b)] + FedNews FedNews t 1

(4.5a) (4.5b)

90-day

Debt markets 3 Year 10 year ANZ CBA

Stock Market NAB

WPC

Index
**

FedSur -0.0047 {0.8122} 0.0129 {0.0089}

*** ***

MacroAnn

0.0008 {0.7668} -0.0041 {0.0225}

**

-0.0022 {0.8659} 0.0003 {0.9651}

-0.1285 {0.3736} 0.0442 {0.7504}

0.0621 {0.3465} -0.0522 {0.7111}

*

-0.1995 {0.2282} -0.1006 {0.4634}

-0.1045 {0.4355} -0.0515 {0.6777}

-0.1630 {0.0189} -0.0917 {0.3164}

FedSur -0.2271 {0.5041} -0.5493 {0.0000}

***

***

***

MacroAnn

-0.2006 {0.0049} -0.1980 {0.0001}

***

-0.1384 {0.0644} 0.2024 {0.0000}

***

USDAUD exchange rates Spot 1M forward 3M forward Conditional mean equations -0.1886 ** -0.1893 ** -0.5017 {0.0243} {0.0426} {0.2320} 0.0580 0.0602 0.0565 {0.5490} {0.5521} {0.5892} Conditional variance equations -0.1878 *** -0.1567 *** 1.0363 *** {0.0000} {0.0000} {0.0002} -0.0659 ** -0.0150 -0.1726 {0.0282} {0.5291} {0.4965} -0.1425 {0.0021} -0.0919 {0.0951}
*

-0.1044 {0.0638} 0.0198 {0.7143}

-0.1726 {0.1314} -0.0381 {0.5769}

-0.1493 {0.1802} -0.0363 {0.5632}

-0.2026 {0.0004} -0.3043 {0.0046}

***

Note: *, **, *** denote significance at 10%, 5%, and 1%, respectively

79

Table 4.9 - Asymmetric spillover effects of the Feds target rate news in the Australian financial markets

This table reports the quasi-maximum likelihood estimates of the EGARCH model as described in equations (4.6a) and (4.6b) of daily Australian stock market returns, interest rate changes, and USDAUD exchange rates. P-values are in braces.

y t = [RHS of (4.4a)] + FedNews _ R FedNews _ Rt 1 + FedSur _ F FedNews _ Ft 1

(4.6a) (4.6b)
WPC
***

ln ht = [RHS of (4.4b)] + FedNews _ R FedNews _ Rt 1 + FedSur _ F FedNews _ Ft 1

90-day

Debt markets 3 Year

10 year

Stock Market NAB

Index

FedSur_P

FedSur_N

***

**

MacroAnn

***

FedSur_P

***

**

***

FedSur_N

MacroAnn

0.0006 {0.1672} 0.0338 {0.0016} -0.0037 {0.0068} -0.4199 {0.0055} 0.1128 {0.6594} -0.1425 {0.4753}

-0.0031 {0.2483} 0.0230 {0.1566} -0.0081 {0.3700} -0.2183 {0.0131} -0.3371 {0.1231} 0.0064 {0.9335}

-0.0049 {0.0379} 0.0194 {0.5150} 0.0136 {0.0004} -1.0555 {0.0014} 0.0427 {0.9548} -0.4799 {0.0002}

-0.2507 {0.5528} 0.4826 {0.3330} -0.0894 {0.5130} -0.1004 {0.4588} -0.3865 {0.0855} -0.0577 {0.3851}
**

-0.2616 {0.0000} 0.5063 {0.1648} -0.0096 {0.9361} -0.0859 {0.3777} -0.4120 {0.0845} -0.0779 {0.4241}

-0.0510 {0.8495} -0.6630 {0.0002} -0.0890 {0.3732} -0.1441 {0.0000} -0.5334 {0.0001} -0.2575 {0.0000}
**

Mean Eq

***

Var Eq

9.7175 {0.0018} 13.0673 {0.0003}

***

2.6698 {0.1023} 0.5413 {0.4619}

0.6720 {0.4124} 1.7840 {0.1817}

USDAUD exchange rates Spot 1M forward 3M forward ANZ CBA Positive and Negative news coefficients ** -0.1289 -0.1277 -0.1332 *** -0.2532 -0.0603 {0.6849} {0.6795} {0.0000} {0.3468} {0.7406} -0.6761 * -0.6381 *** -0.5899 *** 0.6309 0.8849 {0.0723} {0.0013} {0.0000} {0.3329} {0.0453} *** 0.0276 0.0239 0.0160 0.0804 0.0019 {0.7810} {0.8157} {0.8789} {0.6626} {0.9902} *** -0.1162 *** -0.0942 ** -0.1083 ** -0.0980 ** -0.0866 {0.0069} {0.0222} {0.0108} {0.0363} {0.0000} -0.3786 *** -0.2898 *** -0.3617 *** -0.3395 -0.3228 {0.0000} {0.0000} {0.0013} {0.3513} {0.2472} *** -0.0771 *** -0.0369 * -0.0155 -0.1130 * 0.0501 {0.0053} {0.0982} {0.8984} {0.0586} {0.5102} Tests of equality between positive and negative news coefficients 1.2328 1.9224 0.0056 1.4984 3.9053 {0.2669} {0.1656} {0.9403} {0.2209} {0.0481} 9.5106 *** 6.7588 *** 30.1031 *** 0.4218 0.7320 {0.0020} {0.0093} {0.0000} {0.5160} {0.3922}

1.2573 {0.2622} 1.1986 {0.2736}

4.3215 {0.0376} 1.7903 {0.1809}

2.5981 {0.1070} 9.1530 {0.0025}

Note: *, **, *** denote significance at 10%, 5%, and 1%, respectively

80

Table 4.10 - Nonews effects on the Australian financial markets

Panel A - The RBA overall effects with Nonews dummy

This table reports the quasi-maximum likelihood estimates of the EGARCH model as described in equations (4.7a) and (4.7b) of daily Australian stock market returns, interest rate changes, and USDAUD exchange rates. P-values are in braces.
yt = [RHSof (4.3a)] ln ht = [RHSof (4.3b)]+ Nonews Nonewst

(4.7a) (4.7b)

90-day 1.7201 {0.2171} -0.0142 {0.9180}

**

Debt markets 3-Year ANZ 0.9308 {0.3543} -0.0375 {0.7926} CBA

10-year

Stock Market NAB 2.0228 {0.3948} -0.1713 {0.2000}

*

WPC 1.9342 {0.3082} -0.0587 {0.6365}

***

Index -0.0087 {0.9943} -0.1080 {0.2224}

***

***

USDAUD exchange rates Spot 1M Forward 3M Forward Conditional mean equations 0.3728 *** 0.1608 * 0.0295 2.6975 ** 2.2477 ** 1.9236 *** RBASur {0.0000} {0.0911} {0.7788} {0.0388} {0.0276} {0.0092} MacroAnn -0.0032 -0.0007 0.0128 0.0240 0.0297 0.0690 {0.1007} {0.9403} {0.0051} {0.8020} {0.7989} {0.5512} Conditional variance equations RBASur 2.7942 *** 3.6340 *** 1.0406 0.9836 *** 0.6489 *** -1.5988 {0.0000} {0.0000} {0.5435} {0.0000} {0.0011} {0.4016} *** *** MacroAnn -0.6175 0.0326 -0.9810 -0.0524 0.0242 -0.6747 *** {0.0000} {0.3304} {0.0000} {0.1028} {0.3124} {0.0004} 0.7226 *** 0.3433 *** 0.4405 *** -0.0304 * -0.0754 *** 0.6037 *** Nonews {0.0000} {0.0000} {0.0000} {0.0790} {0.0000} {0.0000} Note: *, **, *** denote significance at 10%, 5%, and 1%, respectively 1.9722 {0.0289} -0.1974 {0.0934} 0.1746 {0.1211} 0.5382 {0.4282} 0.0172 {0.8892} -0.0170 {0.8694}

1.0764 {0.0732} -0.0801 {0.2475} 0.0747 {0.0797}

1.7484 {0.0010} -0.0669 {0.2939} 0.0317 {0.5242}

1.7298 {0.0000} -0.3884 {0.0000} 0.1159 {0.0017}

***

81

Table 4.10 Continued

Panel B - The RBA asymmetric effects with Nonews dummy

This table reports the quasi-maximum likelihood estimates of the EGARCH model as described in equations (4.8a) and (4.8b) of daily Australian stock market returns, interest rate changes, and USDAUD exchange rates. P-values are in braces. (4.8a) yt = [RHS of (4.4a)]

ln ht = [RHS of (4.4b)] + Nonews Nonews t

10-year

(4.8b)

90-day

Debt markets 3-Year

Stock Market NAB

WPC

Index

***

***

***

2.2868 {0.5120} -0.1564 {0.9450} -0.1841 {0.1566} 1.0950 {0.1819} *** 0.9468 {0.2823} -0.0814 {0.2404} 0.0755 {0.0765}

1.6409 {0.5285} 2.6607 {0.3954} -0.0527 {0.6705} 2.6808 {0.0000} 0.0080 {0.9963} -0.1031 {0.1068} 0.0469 {0.1855}

-1.0267 {0.4701} 2.1647 {0.3854} -0.0824 {0.3651} 0.4232 {0.5966} 2.7753 {0.0000} -0.3418 {0.0000} 0.0935 {0.0123}

**

USDAUD exchange rates Spot 1M Forward 3M Forward ANZ CBA Positive and Negative news coefficients RBASur_P 0.3870 *** 0.2809 ** 0.1375 ** 3.5427 3.1016 2.8828 *** -0.5470 0.0756 {0.0000} {0.0381} {0.0352} {0.1364} {0.1085} {0.0000} {0.7293} {0.9651} RBASur_N 0.3580 *** -0.0794 -0.1607 0.3029 0.2793 0.1995 3.6590 3.1334 {0.0000} {0.6493} {0.4968} {0.9009} {0.8983} {0.8016} {0.2812} {0.4216} MacroAnn -0.0033 * -0.0044 -0.0040 0.0108 0.0119 0.0229 -0.0114 -0.0047 {0.0850} {0.6943} {0.3631} {0.9096} {0.9035} {0.8146} {0.9330} {0.9760} RBASur_P 1.1775 1.0050 ** -3.2028 1.3952 *** 1.2754 *** -3.0331 2.4182 *** -1.0900 {0.1420} {0.0180} {0.4423} {0.0000} {0.0000} {0.5702} {0.0000} {0.2925} RBASur_N 5.1947 *** 6.3309 *** 2.5826 0.2083 -0.3113 -3.4478 0.8074 2.3374 {0.0000} {0.0000} {0.5178} {0.6086} {0.3522} {0.4244} {0.6081} {0.0022} MacroAnn -0.5556 *** 0.0364 -0.9856 *** -0.0685 ** 0.0010 -0.7092 *** -0.2466 *** 0.0681 {0.0000} {0.2045} {0.0000} {0.0293} {0.9653} {0.0067} {0.0025} {0.3907} 0.7043 *** 0.2923 *** 0.4212 *** -0.0309 * -0.0730 *** 0.6045 *** 0.1974 *** -0.0366 Nonews {0.0000} {0.0000} {0.0000} {0.0589} {0.0000} {0.0000} {0.0004} {0.4936} Tests of equality between positive and negative news coefficients 0.1900 2.6701 1.4777 0.8919 0.9229 6.7228 *** 1.2625 0.5089 Mean Eq {0.6629} {0.1022} {0.2241} {0.3450} {0.3367} {0.0095} {0.2612} {0.4756} 1.0046 6.3383 ** 16.1969 *** 0.0036 0.9319 7.2002 17.2680 *** 96.2578 *** Var Eq {0.0000} {0.0000} {0.3162} {0.0118} {0.0001} {0.9518} {0.3344} {0.0073} Note: *, **, *** denote significance at 10%, 5%, and 1%, respectively

0.2644 {0.6071} *** 0.0152 {0.9019}

0.0632 {0.8015} 2.1612 {0.1415}

1.2361 {0.2662} 6.8520 {0.0089}

***

82

Table 4.10 Continued

Panel C - The Fed effects with the RBAs Nonews dummy

This table reports the quasi-maximum likelihood estimates of the EGARCH model as described in equations (4.9a) and (4.9b) of daily Australian stock market returns, interest rate changes, and USDAUD exchange rates. P-values are in braces.
yt = [RHS of (4.5a)]

(4.9a) (4.9b)

ln ht = [RHS of (4.5b)] + Nonews Nonews t

90-day

Debt markets 3 Year 10 year ANZ CBA

Stock Market NAB

*

WPC

Index
**

FedSur -0.0049 {0.8073} 0.0132 {0.0054}

*** ***

MacroAnn

0.0007 {0.8048} -0.0038 {0.0570}

-0.0019 {0.8959} -0.0006 {0.9489}

-0.1183 {0.3150} 0.0301 {0.8337}

0.0623 {0.3698} -0.0523 {0.7268}

*

-0.1990 {0.0847} -0.1014 {0.4570}

-0.1045 {0.4387} -0.0509 {0.6834}

-0.1653 {0.0140} -0.0936 {0.2894}

FedSur
***

**

***

MacroAnn
***

***

***

***

Nonews

-0.1779 {0.0113} -0.5937 {0.0000} 0.7150 {0.0000}

***

-0.1389 {0.0579} 0.0286 {0.3955} 0.3445 {0.0000}

***

-0.2096 {0.5458} -0.9780 {0.0000} 0.4401 {0.0000}

USDAUD exchange rates Spot 1 m forward 3 m forward Conditional mean equations -0.1886 ** -0.1883 ** -0.4491 {0.0158} {0.0436} {0.2886} 0.0585 0.0614 0.0625 {0.3198} {0.5344} {0.5586} Conditional variance equations -0.1860 *** -0.1528 *** 1.1353 *** {0.0042} {0.0000} {0.0000} -0.0539 *** 0.0116 -0.6923 *** {0.0000} {0.6159} {0.0087} -0.0303 -0.0725 *** 0.6107 *** {0.2820} {0.0000} {0.0000} -0.1400 {0.0075} -0.1630 {0.0021} 0.1469 {0.0001}
***

-0.1045 {0.0711} 0.0291 {0.7525} -0.0187 {0.5052}

-0.1718 {0.1333} -0.0719 {0.2910} 0.0669 {0.1823}

-0.1494 {0.1809} -0.0514 {0.4146} 0.0286 {0.5586}

-0.2033 {0.0001} -0.3492 {0.0000} 0.1142 {0.0006}

***

Note: *, **, *** denote significance at 10%, 5%, and 1%, respectively

83

Table 4.10 Continued

Panel D - The Fed asymmetric effects with the RBAs Nonews dummy

This table reports the quasi-maximum likelihood estimates of the EGARCH model as described in equations (4.10a) and (4.10b) of daily Australian stock market returns, interest rate changes, and USDAUD exchange rates. P-values are in braces.
yt = [RHS of (4.6a)] ln ht = [RHS of (4.6b)] + Nonews Nonewst

(4.10a) (4.10b)

90-day

Debt markets 3 Year 10 year

Stock Market NAB

***

WPC

Index
***

FedSur_P

FedSur_N

***

**

***

MacroAnn

FedSur_P

***

**

***

***

FedSur_N

**

**

***

MacroAnn

***

***

***

Nosurprises

0.0004 {0.7074} 0.0355 {0.0010} -0.0034 {0.0711} -0.3892 {0.0000} 0.1127 {0.4439} -0.5325 {0.0000} 0.7023 {0.0000}

***

-0.0035 {0.2553} 0.0202 {0.1359} -0.0068 {0.5061} -0.2098 {0.0190} -0.3792 {0.0211} -0.0279 {0.7469} 0.2278 {0.0258}

**

-0.0049 {0.0211} 0.0215 {0.4659} 0.0131 {0.0016} -0.9697 {0.0038} 0.0444 {0.9524} -0.9037 {0.0000} 0.4240 {0.0000}

-0.2506 {0.0000} 0.4830 {0.1617} -0.0896 {0.5088} -0.1012 {0.3770} -0.3789 {0.0242} -0.0872 {0.0000} 0.0575 {0.0282}
**

**

-0.2612 {0.3359} 0.5045 {0.2439} -0.0107 {0.9301} -0.0866 {0.3983} -0.4109 {0.1137} -0.0974 {0.1308} 2.8532 {0.4869}
**

-0.0501 {0.0000} -0.6630 {0.0001} -0.0906 {0.4077} -0.1451 {0.0000} -0.5258 {0.0000} -0.2901 {0.0000} 0.0814 {0.0000}

***

Mean Eq

***

***

Var Eq

10.4751 {0.0012} 10.3133 {0.0013} 0.8112 {0.3678} 1.5470 {0.2136}

***

2.9207 {0.0875} 1.8711 {0.1713}

USDAUD exchange rates Spot 1 m forward 3 m forward ANZ CBA Positive and Negative news coefficients ** -0.1290 -0.1279 -0.1282 *** -0.2549 -0.0595 {0.7044} {0.6771} {0.0000} {0.2537} {0.7393} -0.6758 * -0.6301 *** -0.0165 0.6890 0.8905 {0.0817} {0.0017} {0.9802} {0.1255} {0.0422} *** 0.0282 0.0254 0.0326 * 0.0474 0.0016 {0.7725} {0.7993} {0.0734} {0.7291} {0.9918} *** -0.1126 *** -0.0880 ** -2.3120 *** -0.0809 -0.0860 {0.0074} {0.0288} {0.0000} {0.2218} {0.0000} -0.3789 *** -0.2915 *** 2.4336 *** -0.3982 -0.3279 {0.0000} {0.0000} {0.0000} {0.2519} {0.2395} *** -0.0592 ** -0.0086 -0.8323 *** -0.2350 *** 0.0762 {0.0171} {0.6874} {0.0006} {0.0036} {0.3154} *** -0.0444 *** -0.0750 *** 0.6182 *** 0.1815 *** -0.0508 {0.0002} {0.0000} {0.0000} {0.0010} {0.3198} Tests of equality between positive and negative news coefficients 1.0684 1.8696 0.0282 3.5321 * 4.0232 {0.3013} {0.1715} {0.8666} {0.0602} {0.0449} 10.3683 *** 7.7537 *** 66.7768 *** 0.8056 0.7700 {0.0013} {0.0054} {0.0000} {0.3694} {0.3802}

4.4118 {0.0357} 1.8818 {0.1701}

2.2404 {0.1344} 1.3593 {0.2437}

12.4775 {0.0004} 7.9994 {0.0047}

***

Note: *, **, *** denote significance at 10%, 5%, and 1%, respectively

84

CHAPTER 5 - THE SPILLOVER EFFECTS OF TARGET INTEREST RATE NEWS FROM THE US FED AND THE EUROPEAN CENTRAL BANK ON THE ASIA PACIFIC STOCK MARKETS35

35

A shorter version of this chapter is published in Journal of International Financial Markets,

Institutions & Money 19 (3), pp. 415-431, 2009.

85

5.1. Introduction The effect of monetary policy on equity markets has been the subject of interest to not only market participants but also policy markers. A monetary policy shock affects firm investment through its influence on Tobins q, which can conveniently be studied through the response of stock prices to this shock. Simultaneously, stock price reactions have repercussions on consumer behavior through wealth effects. 36 With ample evidence on the global financial linkages, interest rate announcements of major central banks may have spillover impacts on other countries financial markets. Considering the leadership role played by the US markets, any new information on the Feds interest rate policies would have both direct and indirect influences on other stock markets around the world. The indirect influence is via the influence of the US stock market movements resulting from the news announcement on the stock returns in other countries. The literature has provided ample evidence for the information leadership of the US markets, and its influence in the Asia Pacific region in particular, see inter
alia Arshanapalli et al. (1995), Liu et al. (1998), Liu and Pan (1997), Ghosh et al.

(1999), Ng (2000), Miyakoshi (2003), Kim (2003, 2005), Phylaktis and Ravazzolo (2005). Thus, to the extent that the US stock market responds significantly to the Feds news and that the stock markets in other countries look to the U.S markets for

36

For the US Feds news impacts on the US markets, among others, see Bernanke and Blinder (1992),

Kuttner (2001), Bomfim (2003), and Bernanke and Kuttner (2005).

86

trading momentum, there is potential for the Fed news to be significantly priced in other markets. The direct influence of the Feds interest rate news can be examined by investigating the extent to which markets in other countries respond to the news. The spillover effects of the Feds news have been a topic of interest not only to policy makers but also to market participants alike. Thus far, there are only a handful of studies examining the direct influence of the Feds news on the financial markets of other countries. Wongswan (2006, 2009) provide mixed evidence on the Feds interest news spillover effects on the stock market returns in a number of countries. Ehrmann and Fratzscher (2006) report that the US monetary policy news has a negative impact on the returns of fifty equity markets including the twelve Asia Pacific markets examined in this chapter. However, none of these papers examines the Feds target interest rate news spillover effects on the volatility of market returns in other countries.37 This is a significant oversight because the news does not only affect the direction of market movements in these countries (first moment influence) but also influences the trading environment, and hence the level of volatility (second moment effect). Another oversight in the literature is the lack of investigation of the spillover effects of the ECBs interest rate news. Given the increasing prominence of the euro areas equity markets, there is a strong potential for spillover impacts of the ECBs news in the stock markets of other regions. In particular, considering the growing bilateral flows of trade and financial investments between the Euro area and

37

However, there is limited evidence of the Fed target rate surprises increasing the Irish equity market

volatility (see Bredin et al., 2005).

87

the Asia Pacific region, 38 the ECBs monetary policy decisions are keenly monitored by the market participants in the latter. In addition, the literature is silent on how quickly the news from these two central banks is absorbed in foreign stock markets. This chapter aims to address these important shortfalls in the literature and provide comprehensive evidence on the existence and the nature of the spillover effects of the Feds and the ECBs policy interest rate news in the Asia Pacific region. In particular, this chapter investigates the direct news effects on the first two moments of the stock market returns in twelve Asia Pacific economies for the period January 1999 to December 2006. This study contributes to the literature in many dimensions. First, extensive evidence of the direct effects of the US Feds and the ECBs target rate news in the Asia Pacific region has been provided. Not only the presence of the news effects has been documented, but also additional contributions of the target rate news that are significant over and above the documented spillover effect from the US and euro areas stock markets have been provided. In addition, potential of heterogeneity of the directions of the news impacts and the varying degrees of strengths of responses across the twelve countries has also been investigated. Second, comprehensive evidence of the news effects on the second moments of stock returns has been provided to see if the news significantly increase or decrease volatility in the Asia Pacific region. Third, the target rate news effects have been investigated across both overnight and intradaily holding periods of the

38

The EU is the largest trading partner for Australia and China; second largest trading partner for

Hong Kong, New Zealand, Taiwan, and Thailand; the third for Indonesia, Japan, Malaysia, and Singapore; and the fourth for Korea and the Philippines (see http://ec.europa.eu/trade/issues/bilateral/ data.htm).

88

Asia Pacific markets. This enables investigations of potential differences in the speeds of the news absorption across the region. In short, this is the first research that reports such comprehensive target interest rate news spillovers effects from the US Fed and the ECB on the Asia Pacific markets in terms of not only establishing the first and second moment effects but also detailing the country specific responses in (i) the directions and the strengths of the news effects and (ii) the varying degrees of speed of market adjustments to the news. The key findings of this chapter are summarized as follows. First, in general, the Asia Pacific stock market returns fell (rose) significantly in response to unexpected interest rate rises (cuts) from both the Fed and the ECB, which is largely consistent with the previous literature. Second, the interest rate news from the two central banks significantly raised the return volatility in most of the markets. Third, both the Fed and the ECB news were absorbed gradually leading to delayed or persisting volatility responses. These results are robust when controlling for the joint impact of the Feds and the ECBs policy rate news. The rest of the chapter is organized as follows. Section 5.2 discusses data and empirical modeling issues, Section 5.3 reports and analyzes the estimation results. Robustness checks and other tests are presented in Section 5.4, and Section 5.5 concludes the chapter. 5.2. Data and empirical modeling issues 5.2.1. The Asia Pacific stock index returns While most of the previous studies utilize daily closing data to explore the impacts of the target rate news, this study employs opening and closing prices to

89

disaggregate the daily investigation horizon into overnight and intraday periods. Daily open and close prices were obtained from Bloomberg for the stock indices of twelve Asia Pacific countries: Australia, China, Hong Kong, Indonesia, Japan, Korea, Malaysia, New Zealand, the Philippines, Singapore, Taiwan, and Thailand for the period January 1999 to December 2006. Specifically, the stock indices examined are Standard & Poors/Australian Stock Exchange 200 Index (S&P/ASX 200), Shanghai Stock Exchange Composite Index (SHCOMP), Hang Seng Index (HSI), Jakarta Stock Exchange Composite Index (JCI), NIKKEI 225 Index (NKY), Korea Exchange 100 Index (KRX100), Kuala Lumpur Stock Exchange Composite Index (KLCI), New Zealand Exchange Limited 50 Free Float Total Return Index (NZSE50FG), Philippine Stock Exchange PSEi Index (PCOMP), Straits Times Index (STI), Taiwan TWSE Index (TWSE), and Stock Exchange of Thai Index (SET).39 Hamao et al. (1990) suggest that information spillovers from foreign markets would be expected to show up only in the close to open returns as predicted by international asset pricing model. On the other hand, volatility spillovers to the conditional variances of the close-to-open and open-to-close horizons can occur (pp. 282-3). However, this chapter hypothesizes that it is possible to have significant responses during the open-to-close horizon if there are delayed responses to the overnight interest rate news from the Fed and the ECB. It can be argued that if the interest rate signal is clear and the individual countrys central bank reaction is fairly predictable, then it is expected to find the spillover influence only on the mean

39

Due to data limitations, daily open and close data for New Zealand and Korea were obtained from

January 03, 2001, and from April 27, 2004 for Indonesia.

90

returns during the close to open horizon as Hamao et al. (1990) suggest. However, if market participants are unsure about their authoritys monetary policy responses (whether to follow the overnight leads of the Fed and the ECB and adjust their policy rates in the same direction), then this indecision may be shown in both the conditional mean and the variance of the open-to-close returns. Figure 5.1 - Testing horizons H2 H1 H0 Close Day t-1 Open Day t Close

Therefore, stock index returns have been calculated over three time horizons. The first horizon (H0) is over the close on calendar day t-1 to the close on calendar day t (close-to-close); the overnight horizon (H1) is over the close of day t-1 to the open on day t (close-to-open); the intraday horizon (H2) is measured over the open on day t to the close on day t (open-to-close). Figure 5.1 visualizes these three time horizons. Since both the US Feds and the ECBs target interest rate announcements were made when the Asia Pacific markets were closed, the earliest market reaction on the news can be captured at market opening one calendar day after the news announcement. Thus, the overnight horizon (H1) captures the Asia Pacific stock markets first reaction to the news, whereas the intradaily horizon (H2) captures any delayed reactions during the trading day in the Asia Pacific region. Table 5.1 reports the summary statistics of the stock market returns across the three horizons. There are significant differences across the three time horizons in various aspects.
91

Table 5.1 - Descriptive statistics of the Asia Pacific stock markets returns

This table reports descriptive statistics for the Asia Pacific stock index returns series for the period from January 1999 to December 2006. P-values are in braces.

H0 2084 0.0352 0.7200 -0.5415 3.7404 1316.70 698 0.1144 1.2318 -0.9368 5.8670 1103.21
*** ** *** *** ***

Australia H1 2084 0.0421 0.8060 -0.2841 3.1327 880.21 H2 2084 -0.0071 0.3507 -0.0425 11.2917 11077.42 H0 2084 0.0387 1.3588 0.5900 5.6265 2869.85 H2 2084 0.0182 0.5659 -3.2757 57.2628 288593.99 H0 2084 0.0330 1.3268 -0.2728 3.6180 1162.47 H2 2084 -0.0297 0.7740 0.3994 6.5412 3772.56 H0

China H1 2084 0.0206 1.5893 0.9240 7.3291 4960.84

Hongkong H1 2084 0.0631 1.5508 -0.0823 2.2099 426.41

Indonesia H1 698 0.0861 1.2698 -1.1398 5.7333 1107.11

***

H2 698 0.0282 0.4558 1.6434 17.9737 9723.56

*

***

No. of observations Mean Std. deviation Skewness Kurtosis Jarque-Bera Ljung-Box Q test Return P-Value Volatility P-Value 83.8408 {0.0000} 2838.2129 {0.0000}
*** *** *** *** *** ***

20.1589 {0.4480} 241.7752 {0.0000}

***

112.4802 {0.0000} 487.0293 {0.0000}

***

35.1171 {0.0195} 237.5584 {0.0000}

79.5264 {0.0000} 234.8783 {0.0000}

719.7540 {0.0000} 117.1691 {0.0000}

21.6317 {0.3608} 340.1265 {0.0000}

209.8833 {0.0000} 503.7688 {0.0000}

25.0182 {0.2007} 189.0722 {0.0000}

***

25.8521 {0.1708} 155.7645 {0.0000}

***

43.9239 {0.0015} 156.0778 {0.0000}

***

29.7625 {0.0738} 58.4805 {0.0000}

***

H0 2084 0.0120 1.3512 -0.1171 1.8822 312.37

** *** *** ***

Japan H1 2084 0.0574 1.4398 -0.0824 1.0871 104.97 H2 2084 -0.0455 0.5803 0.0590 1.7939 280.79 H0 1562 0.0695 1.6552 -0.4191 4.1435 1163.10 H2 1562 -0.1034 1.8129 0.4165 2.7251 528.81 H0 2084 0.0297 1.0175 -0.1592 6.4206 3588.42

Korea H1 1562 0.1731 2.4608 -0.3460 1.9268 272.80

Malaysia H1 2084 0.0514 1.1030 0.4549 9.9974 8750.68

***

H2 2084 -0.0209 0.3805 -1.1818 58.1411 294156.58

***

H0 1562 0.0565 0.6725 -0.6334 5.4663 2049.16

***

New Zealand H1 1562 0.0651 0.8347 -0.2064 5.2015 1771.98

***

H2 1562 -0.0097 0.5191 0.9222 17.5585 20299.70

***

***

No. of observations Mean Std. deviation Skewness Kurtosis Jarque-Bera Ljung-Box Q test Return P-Value Volatility P-Value 37.3035 {0.0108} 127.8315 {0.0000}
*** *** *** ***

14.1155 {0.8246} 306.3368 {0.0000}

***

79.6269 {0.0000} 253.6935 {0.0000}

***

23.2339 {0.2775} 120.2963 {0.0000}

496.3204 {0.0000} 354.3115 {0.0000}

84.9173 {0.0000} 203.2037 {0.0000}

96.8672 {0.0000} 594.8878 {0.0000}

***

154.5961 {0.0000} 516.0267 {0.0000}

***

105.7054 {0.0000} 57.5496 {0.0000}

***

38.3396 {0.0080} 276.9891 {0.0000}

***

300.2574 {0.0000} 943.5128 {0.0000}

***

84.5225 {0.0000} 636.6785 {0.0000}

***

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively.

92

Table 5.1 Continued

H2 2084 -0.0124 0.4869 -5.0148 118.6021 1230761.57
*** *** *** *** *** *** ***

H0 2084 0.0190 1.2974 1.5492 21.6387 41491.76

Philippines H1 2084 0.0313 1.3884 1.4065 15.8593 22527.08 H0 2084 0.0362 1.1634 -0.3801 4.8030 2053.33 H2 2084 -0.0632 0.6099 1.4045 22.1588 43342.38 H0 2084 0.0089 1.5668 -0.0780 3.0127 790.23 H2 2084 -0.1083 0.9434 0.7083 5.1557 2483.56 H0 2084 0.0310 1.5411 -0.1760 9.8920 8507.47

Singapore H1 2084 0.0992 1.3432 -0.0058 3.9919 1383.73

Taiwan H1 2084 0.1181 1.9545 -0.4690 3.2690 1004.31

Thailand H1 2084 0.0939 1.6144 -0.1276 7.8058 5296.39

***

H2 2084 -0.0627 0.6260 3.3490 45.8782 186753.05

***

***

No. of observations Mean Std. deviation Skewness Kurtosis Jarque-Bera Ljung-Box Q test Return P-Value Volatility P-Value
***

50.9570 {0.0002} 15.7768 {0.7304}

** *** *** *** *** ***

173.6484 {0.0000} 250.4691 {0.0000}

40.2383 {0.0047} 32.3908 {0.0393}

27.7714 {0.1149} 291.1947 {0.0000}

161.7326 {0.0000} 517.2748 {0.0000}

28.0720 {0.1077} 50.8729 {0.0002}

44.0922 {0.0015} 569.8466 {0.0000}

179.1819 {0.0000} 776.3626 {0.0000}

96.7493 {0.0000} 542.9126 {0.0000}

***

41.7871 {0.0029} 184.6199 {0.0000}

***

75.2960 {0.0000} 215.9278 {0.0000}

***

49.1880 {0.0003} 73.0767 {0.0000}

***

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively.

93

For the mean, in most cases the series have a positive mean for the daily (H0) and overnight (H1) horizons, and a negative mean for the intradaily (H2) horizon. In addition, the returns are generally higher during overnight than intradaily. In all cases, the return volatility as measured by the standard errors is higher during overnight than during the other horizons and this is consistent with the literature. In most cases, stock returns are negatively skewed over daily and overnight, and positively skewed over the intradaily horizon. The reverse is observed for the Philippines and China. The Australian stock market returns show a negative skewness over all three horizons. In most cases, the returns exhibit leptokurtosis with the highest excess kurtosis over intradaily except for Japan. Furthermore, the returns show significant serial correlation in the second moments over all three time horizons. 5.2.2. Empirical modelling We start with the baseline univariate EGARCH(1,1) model and then progress to specific modeling of the impacts of the Feds and the ECBs news on the Asia Pacific stock markets. The baseline EGARCH (1,1) model employed in our study is described by the conditional mean and the conditional variance equations (5.1a) and (5.1b) shown below.
y t = c + Lag yt 1 + Hol Holt + t
ln ht = c + h ln ht 1 + 1

t 1
ht 1

+ 2

t 1
ht 1

+ Hol Holt

(5.1a) (5.1b)

The conditional mean equation for the returns ( yt ) is expressed as a function of a one day lag of returns and a holiday dummy in relevant markets. The holiday
94

dummy is used to account for the potential impacts of differences in trading days across stock markets due to market closure. It is assigned to each stock market and takes the number of days of market closure between two successive market prices. For example, for normal consecutive daily observations (i.e. returns calculated over two days - Monday close to Tuesday close price) the value of zero is assigned, whereas values of one or higher will be assigned for returns observations calculated over a longer horizon due to market closure. This seasonal dummy helps to control for the days of more intense information flows following a longer period of market closure. The conditional variance equation for the returns in the financial market series ( ht ) is expressed as a function of one period lags of the variance and the residuals, and the holiday dummy. One lag of the dependant variable has been chosen as this addressed the residual serial correlations in most cases. To investigate the spillover impacts of the Feds and the ECBs news on the Asia Pacific stock market returns and volatilities, then the baseline model has been extended with a news variable for each of the Fed and the ECB, Newst (FedNewst for the Fed, and ECBNewst for the ECB estimated separately). The news variable is lagged by one period to account for the time difference between the US and European trading hours and the Asia Pacific trading hours. In addition, in the course of a particular trading day, macroeconomic announcements from the USs and the euro areas agencies, other than monetary policy decisions, might hit the markets. Thus, Asia Pacific stock markets would also be influenced by these macro announcements in addition to the target interest rate news when they open for trading after the close of the US and the euro areas stock markets.

95

Furthermore, as all macroeconomic announcements of the US and the euro area are released before the Feds and the ECBs target rate announcements, 40 a dummy variable (MacroAnni,t) for each of the major macroeconomic announcements has been added to the conditional variance equation in order to isolate the influence of the target interest rate news on these days of multiple information releases. For the US, the following announcements have been controlled: Gross Domestic Product growth rate (GDP), Consumer Price Index (CPI), Unemployment rate (UE), Trade deficit (TD), and Retail sales growth rate (RET). For the ECBs target rate news estimations, the following euro area announcements have been controlled: Gross Domestic Product growth rate (GDP), Consumer Price Index (CPI), Unemployment rate (UE), External trade (ET), Current account balance (CAB), and Retail trade growth rate (RTT). 41 Among the total of 69 and 131 target rates announcements made by the Fed and the ECB, there were 13 and 27 macroeconomic announcements released by the USs and the euro areas agencies on the same dates. This represents 18.84% and 20.61% of the Feds and the ECBs announcement sample, respectively (see Appendix D). The extended model is then shown in equations (5.2a) and (5.2b) below.
y t = [RHS of (5.1a)] + News News t 1

(5.2a) (5.2b)

ln ht = [RHS of (5.1b)] + News Newst 1 + MacroAnn,i MacroAnni ,t 1

i

40

The USs macroeconomic announcements are normally released at 8:30 AM EST, while the euro

areas macroeconomic announcements are made at 11:00 CET.

41

Due to data limitation, data on the euro areas macroeconomic announcements was only available

since January 2002 onwards.

96

In general, the target news coefficient in the conditional mean equation is expected to be negative as Wongswan (2006, 2009), Bernanke and Kuttner (2005), Ehrmann and Fratzscher (2006), and Rigobon and Sack (2004) documented for various stock markets. Unexpected interest rate rises from the Fed and the ECB could have similar market depressing influences in the Asia Pacific region (a negative news coefficient) at least through three factors: (i) higher financing costs for foreign companies that raised capital in the US/euro area markets; (ii) lower stock prices for foreign companies listed in the US/the euro area markets or for those having operations or links with the US/the euro area; and (iii) lower demand from the US/the euro area for foreign commodities (consumption effect). The news effect on the conditional volatilities would depend on whether the news adds to or resolves uncertainties in the market. As our volatility specification allows negative coefficients in the conditional volatility equation, it is possible to have both positive and negative sign for the interest rate news coefficient, News. In the days approaching the next announcement, the market may settle towards some degree of homogeneity of beliefs. If an unexpected change in the target interest rate is announced, the homogeneity evaporates resulting in heightened heterogeneity. This in turn would lead to excited transaction volumes and thus, conditional price volatility. If that is the case, one would observe a positive News as Wongswan (2009) report for a volatility rising effects of the Feds interest rate news. On the other hand, if an unexpected rate change announcement resolves the accumulated heterogeneity in the market and hence reduces the level of uncertainty, a significant volatility reduction would be observed (i.e. a negative News).

97

In addition, to the extent that the Feds rate decisions can influence those of the ECB, it is possible that the impacts of the ECBs rate decisions on Asia Pacific stock markets will be weaker as the market participants might regard the ECBs news as less newsworthy. Granger causality tests as shown in Table 5.2 revealed that there is a strong evidence of the Feds FOMC Granger causing the ECBs Governing Council in making interest rate decisions, while the evidence for the reverse is weak. Table 5.2 - Granger causality tests for the Fed and the ECB
This table reports the Granger causality tests results for the hypotheses that the Fed Granger causes the ECB in target rates decision making process (the left panel), and the ECB Granger causes the Fed in target rates decision making process (the right panel). P-values are in braces.

Fed Granger causes ECB ECB Granger causes Fed Lags F-value Lags F-value 1 0.7528 1 0.000002 {0.3857} {0.9987} 2 6.4494 *** 2 0.000002 {0.0016} {0.999998} 3 2.2636 3 4.2934 *** {0.0050} {0.7920} 4 3.2162 ** 4 1.6959 {0.0121} {0.1482} 5 2.5774 ** 5 1.3628 {0.0248} {0.2353} Note: *, **, *** denote significance at 10%, 5%, and 1%, respectively

Therefore, it is necessary to control for the other central banks news effect to check for the robustness of the target rate news impacts, especially in the case of the ECBs news investigation. Equations (5.3a) and (5.3b) below include both the Feds and ECBs target news coefficients (FedNewst, and ECBNewst) and both the USs and the euro areas macroeconomic announcements dummies.
y t = [RHS of (5.1a)] + FedNews FedNews t 1 + ECBNews ECBNews t 1

(5.3a)

ln ht = [RHS of (5.1b)] + FedNews FedNewst 1 + USMacroAnn,iUSMacroAnni ,t 1

i

+ ECBNews ECBNewst 1 + EuroareaMacroAnn,i EuroareaMacroAnni ,t 1

i

(5.3b)

98

The ECBs news coefficient on the mean would be insignificant if the ECBs news is stale when it hits the Asia Pacific markets. However, if the ECBs news is also relevant, news coefficient is expected to be significant. On the variance equation, the impact of the ECBs news would once again depend on its supplemental role for the Feds news in either raising or lowering the level of market uncertainty in the Asia Pacific region. 5.3. Empirical results 5.3.1. Baseline model results Table 5.3 shows the baseline models results. In general, the Holiday dummy tends to reduce returns over the overall and intradaily horizons while it has a positive effect during overnight horizon. In the conditional variance equations, the lagged variance term (h) is close to one in all cases suggesting volatility persistence as found in the literature. There is evidence for an asymmetric influence of the past innovations as shown by a negative 1, in all cases for the overnight (H1) and the daily (H0) horizons. This is consistent with the common findings in the literature. The volume effect is positive in all cases (2>0) indicating that regardless of sign, the higher is the previous periods unexpected stock movements, the higher its impact on the conditional variance in the next period. The conditional variances are generally higher on the days immediately following market closure due to holidays. The serial correlations in the first and the second moments of the standardized residuals are removed in all cases by adding more than one lags of the dependent variable in the mean equation and by varying the lag structures of the EGARCH model in the variance equation. However, the news coefficients from these estimations are qualitatively the same as the ones reported in this chapter. Thus, for
99

consistency, the same model specification has been adopted for all twelve stock indices and the results are reported accordingly. 5.3.2. The US Feds target interest rate news spillover effects The quasi-maximum likelihood estimates of the EGARCH model for the spillover effects of the Feds target rate surprises as modeled in (5.2a) and (5.2b) are reported in Table 5.4. The impacts of the spillover news effects on the Asia Pacific financial markets are investigated by examining the sign and the magnitude of the news variables in each estimation over different time horizons. Consistent with the findings in the literature, this study finds that unexpected hike in the Feds target rate significantly reduced the Asia Pacific stock market returns, in general. Eight out of twelve stock markets show significant responses to the news and the news coefficient is negative in all cases. This may be due to the impact of the Feds unexpected hike on the financing costs for foreign borrowers, and/or the consumption effects spillover from the US economy to other economies. Ultimately, these adverse impacts have been reflected in significantly lower returns in the Asia Pacific stock markets. On the other hand, there is mixed evidence on the speed of the Fed news absorption. Some markets responded quickly as evidenced by significant news coefficient in overnight horizon only (Hong Kong and Singapore), while in one market (Taiwan) the news effect shows up only during intradaily horizon. There is a possibility that in some markets an opening value of the index may not always incorporate active opening prices of all stocks constitute the index due to delays of some stocks in generating opening prices. This may have contributed to the significant responses during the intradaily horizon. In three other markets (China, Malaysia and Thailand), the new effects are significant during both overnight and intradaily horizons.

100

Table 5.3 - Baseline model results

This table reports the estimation results of the EGARCH models described in (5.1a) and (5.1b) for Asia Pacific stock markets daily returns. H0 is daily horizon over the close on calendar day t-1 to the close on calendar day t; H1 is overnight horizon over the close of day t-1 to the open on day t; and H2 is intraday horizon measured over the open on day t to the close on day t. P-values are in braces.

y t = c + Lag yt 1 + Hol Holt + t

ln ht = c + h ln ht 1 + 1 ht 1
Hongkong H1 H2

t 1
+ 2 ht 1 + Hol Holt

t 1

(5.1a) (5.1b)

H0

Australia H1 H0

H2

H0

Indonesia H1

H2

Lag, 1

Holiday

0.0316 *** 0.0243 * 0.0179 *** {0.0063} {0.0597} {0.0000} -0.0079 0.1750 *** 0.1151 *** {0.2787} {0.0000} {0.0000} 0.0233 * 0.0144 0.0003 {0.0782} {0.6952} {0.9853}

0.0616 *** {0.0000} 0.0132 {0.5002} 0.0349 {0.6134}

0.0495 ** -0.0317 ** {0.0240} {0.0184} 0.3119 *** 0.0186 {0.0000} {0.3443} 0.1675 ** -0.1271 ** {0.0178} {0.0487}

0.1085 *** 0.0827 *** 0.0241 *** {0.0000} {0.0086} {0.0000} 0.1629 *** 0.2455 *** 0.0062 {0.0000} {0.0000} {0.8161} -0.0085 -0.0041 0.0241 {0.5712} {0.8550} {0.2167}

Holiday

0.6857 *** {0.0000} 0.9681 *** {0.0000} 0.1139 *** {0.0000} 0.1095 *** {0.0000} 0.0543 * {0.0797}

0.9782 *** {0.0000} 0.9787 *** {0.0000} 0.0963 *** {0.0000} 0.1385 *** {0.0000} 0.1004 ** {0.0448}

0.5987 *** {0.0000} 0.9824 *** {0.0000} -0.0360 *** {0.0000} 0.2039 *** {0.0000} 0.0013 {0.9439}

-0.0888 *** {0.0000} 0.9733 *** {0.0000} -0.0262 *** {0.0000} 0.1437 *** {0.0000} 0.2664 *** {0.0000}

China H1 H2 H0 Conditional mean equation -0.0244 0.0360 *** -0.0240 *** {0.3739} {0.0000} {0.0000} 0.1290 *** 0.2066 *** 0.0185 ** {0.0000} {0.0000} {0.0114} -0.2203 ** 0.0342 -0.0360 *** {0.0126} {0.7016} {0.0029} Conditional variance equation -0.1621 *** -0.4578 *** -0.0633 *** {0.0000} {0.0000} {0.0000} 0.9575 *** 0.8556 *** 0.9930 *** {0.0000} {0.0000} {0.0021} -0.0472 *** 0.2500 *** -0.0255 *** {0.0001} {0.0000} {0.0000} 0.2625 *** 0.5097 *** 0.0879 {0.0000} {0.0000} {0.6333} 0.0918 *** 0.1876 *** 0.0808 {0.0048} {0.0014} {0.3006} -0.0886 *** {0.0000} 0.9870 *** {0.0000} -0.0375 *** {0.0001} 0.1240 *** {0.0000} 0.1461 *** {0.0065} -0.1672 *** {0.0000} 0.9606 *** {0.0000} 0.0859 *** {0.0000} 0.1903 *** {0.0000} 0.0823 {0.1755}

-0.1782 *** {0.0000} 0.8796 *** {0.0000} -0.1636 *** {0.0000} 0.2648 *** {0.0000} -0.1170 {0.7987}

-0.1569 *** {0.0000} 0.8648 *** {0.0000} -0.1744 *** {0.0000} 0.2419 *** {0.0000} -0.2590 {0.8088}

-0.1664 *** {0.0000} 0.9646 *** {0.0000} -0.0584 *** {0.0000} 0.1748 *** {0.0000} -0.3822 *** {0.0000}

101

Table 5.3 Continued

LogL Q(20)

Q (20)

-198 12.7711 {0.8870} 17.2503

-343 2431 19.0121 27.7282 {0.5210} {0.1160} 37.1101 ** 14.4536

-3449 33.0896 ** {0.0330} 9.5363

Estimation diagnostics -3689 -1273 -3330 33.7848 ** 49.6201 *** 16.1528 {0.0276} {0.0003} {0.7071} 11.9885 0.9388 29.2455 * -3591 -2233 42.0562 *** 14.4936 {0.0027} {0.8046} 22.6971 11.4919 -1050 11.0269 {0.9455} 21.9599

-1058 14.9888 {0.7770} 31.0980 *

-358 18.7718 {0.5367} 9.2140

{0.6367} {0.0268} {0.8068} {0.9758} {0.9165} {1.0000} {0.0830} {0.3039} {0.9325} {0.3427} {0.0539} {0.9803} Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively. Q(20) and Q2(20) are Ljung-Box portmanteau test of white noise for linear and non-linear (squared) standardized residual.

102

Table 5.3 - Continued

H0 Malaysia H1 H2 H0 New Zealand H1 H2

H0

Japan H1

H2

Lag, 1

Holiday

0.0256 {0.2367} -0.0048 {0.8502} -0.0215 {0.9403}

0.0276 -0.0283 ** {0.4048} {0.0146} 0.1661 *** 0.0532 ** {0.0000} {0.0132} 0.2020 *** -0.1262 *** {0.0000} {0.0000}

0.0854 *** {0.0000} -0.0088 {0.7345} 0.0349 {0.6134}

0.0159 0.0050 *** {0.2558} {0.0000} 0.2474 *** 0.1434 *** {0.0000} {0.0000} 0.0224 -0.0353 {0.4587} {0.5416}

0.0572 *** 0.0522 *** 0.0055 *** {0.0000} {0.0001} {0.0000} 0.0728 *** 0.2709 *** 0.0785 *** {0.0022} {0.0000} {0.0000} -0.0085 -0.0041 0.0241 {0.5712} {0.8550} {0.2167}

Holiday

-0.1001 *** {0.0000} 0.9741 *** {0.0000} -0.0606 *** {0.0000} 0.1486 *** {0.0000} -0.0775 *** {0.0007}

-0.6175 *** {0.0000} 0.8192 *** {0.0000} -0.0223 *** {0.0001} 0.1007 *** {0.0000} -0.9914 *** {0.0000}

-0.2024 *** {0.0000} 0.9484 *** {0.0000} -0.0002 {0.9889} 0.1820 *** {0.0000} -0.1645 *** {0.0000}

-0.0754 *** {0.0000} 0.9785 *** {0.0000} -0.0606 *** {0.0000} 0.1298 *** {0.0000} 0.2664 *** {0.0000}

-0.1314 *** {0.0000} 0.9880 *** {0.0000} -0.0392 *** {0.0000} 0.1758 *** {0.0000} 0.2391 *** {0.0007}

-0.1422 *** {0.0000} 0.9860 *** {0.0000} -0.0577 *** {0.0000} 0.2065 *** {0.0000} -0.1728 {0.3440}

-0.1129 *** {0.0000} 0.9864 *** {0.0000} -0.0317 *** {0.0000} 0.1302 *** {0.0000} -0.1170 {0.7987}

-0.1042 *** {0.0000} 0.9952 *** {0.0000} -0.0119 {0.4303} 0.1290 *** {0.0000} -0.2590 {0.8088}

-0.1541 *** {0.0000} 0.9706 *** {0.0000} 0.0146 {0.6021} 0.1604 *** {0.0000} -0.3822 *** {0.0000}

LogL Q(20) -2881 15.9978 {0.7168} 25.1324

Q (20)

-3464 9.4727 {0.9768} 28.6844 *

-3586 14.3928 {0.8100} 18.7070

-1716 26.2666 {0.1572} 7.6719

Korea H1 H2 H0 Conditional mean equation 0.0628 -0.0761 ** 0.0277 ** {0.1142} {0.0435} {0.0181} 0.5439 *** 0.2497 *** 0.1605 *** {0.0000} {0.0000} {0.0000} -0.2203 ** 0.0342 -0.0377 {0.0126} {0.7016} {0.6950} Conditional variance equation -0.1010 *** -0.0914 *** -0.1111 *** {0.0000} {0.0000} {0.0012} 0.9701 *** 0.9672 *** 0.9903 *** {0.0000} {0.0000} {0.0000} -0.0918 *** 0.0760 *** -0.0370 ** {0.0000} {0.0000} {0.0111} 0.1787 *** 0.1615 *** 0.1469 *** {0.0000} {0.0000} {0.0008} 0.0918 *** 0.1876 *** 0.2271 *** {0.0048} {0.0014} {0.0000} Estimation diagnostics -3229 -3007 -2581 136.7139 *** 19.7791 27.2055 {0.0000} {0.4718} {0.1296} 25.3701 14.8529 39.7823 *** -2664 24.2447 {0.2319} 29.1460 * -495 15.2344 {0.7628} 4.2625 -1463 19.1492 {0.5122} 9.7853

-1588 82 32.0719 ** 18.0648 {0.0425} {0.5831} 18.9662 2.4437

{0.0942} {0.5409} {0.9938} {0.1964} {0.1876} {0.7848} {0.0053} {0.0849} {0.9999} {0.9719} {0.5240} {1.0000} Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively. Q(20) and Q2(20) are Ljung-Box portmanteau test of white noise for linear and non-linear (squared) standardized residual.

103

Table 5.3 Continued

H0 Taiwan H1 H2 H0 Thailand H1 H2

H0

Philippines H1

H2

Lag, 1

Holiday

0.0144 *** 0.0240 -0.0157 *** {0.0009} {0.2495} {0.0000} 0.4266 *** 0.2472 *** 0.0390 ** {0.0000} {0.0000} {0.0298} -0.2310 *** 0.0622 *** -0.2353 *** {0.0000} {0.0000} {0.0000}

0.0480 *** {0.0037} 0.0283 {0.1880} 0.0706 * {0.0502}

0.1264 *** -0.1313 *** {0.0000} {0.0000} 0.2663 *** 0.1013 *** {0.0000} {0.0005} 0.1159 *** -0.1411 ** {0.0085} {0.0167}

0.0583 *** 0.0743 *** -0.0805 *** {0.0006} {0.0012} {0.0000} 0.0644 *** 0.1846 *** 0.0540 *** {0.0063} {0.0000} {0.0020} -0.1596 *** -0.0647 -0.0816 *** {0.0001} {0.6902} {0.0008}

Holiday

-0.0617 *** {0.0000} 0.9324 *** {0.0000} -0.0330 *** {0.0000} 0.1297 *** {0.0000} 0.6981 *** {0.0000}

-0.2276 *** {0.0000} 0.8443 *** {0.0000} -0.0238 {0.1294} 0.3870 *** {0.0000} 0.3177 *** {0.0036}

-0.5872 *** {0.0000} 0.7354 *** {0.0000} 0.2187 *** {0.0000} 0.2818 *** {0.0000} 0.7197 *** {0.0000}

-0.1510 *** {0.0000} 0.9801 *** {0.0000} -0.0421 *** {0.0000} 0.2021 *** {0.0000} 0.0357 {0.6995}

-0.1336 *** {0.0000} 0.9743 *** {0.0000} -0.0669 *** {0.0000} 0.2035 *** {0.0000} 0.0997 *** {0.0085}

-0.1961 *** {0.0000} 0.9647 *** {0.0000} 0.0403 *** {0.0003} 0.2428 *** {0.0000} 0.0784 ** {0.0171}

-0.0988 *** {0.0000} 0.9265 *** {0.0000} -0.0719 *** {0.0000} 0.1733 *** {0.0000} 0.4529 *** {0.0000}

-0.1317 *** {0.0000} 0.8799 *** {0.0000} -0.0941 *** {0.0000} 0.3119 *** {0.0000} 0.4849 *** {0.0000}

-0.3204 *** {0.0000} 0.8891 *** {0.0000} -0.0328 *** {0.0000} 0.3258 *** {0.0000} 0.4609 *** {0.0000}

LogL Q(20) -3041 19.1775 {0.5103} 10.7515

Q (20)

-3418 15.2965 {0.7592} 2.2852

-3349 31.2546 * {0.0519} 24.6351

-1183 19.3104 {0.5017} 7.4552

Singapore H1 H2 H0 Conditional mean equation 0.0927 *** -0.0858 *** 0.0308 ** {0.0000} {0.0000} {0.0491} 0.2267 *** 0.0179 0.0509 ** {0.0000} {0.5788} {0.0166} 0.0424 -0.0518 *** 0.0203 {0.4741} {0.0003} {0.1639} Conditional variance equation -0.1768 *** -0.2847 *** -0.1115 *** {0.0000} {0.0000} {0.0000} 0.9778 *** 0.9243 *** 0.9817 *** {0.0000} {0.0000} {0.0000} -0.0472 *** 0.1602 *** -0.0693 *** {0.0000} {0.0000} {0.0000} 0.2360 *** 0.2653 *** 0.1537 *** {0.0000} {0.0000} {0.0000} 0.1461 ** -0.0731 0.1375 *** {0.0217} {0.6163} {0.0000} Estimation diagnostics -3244 -1676 -3644 40.2383 *** 22.0282 21.9908 {0.0047} {0.3390} {0.3410} 17.2090 4.5056 26.9474 -3984 24.0788 {0.2390} 23.8491 -2518 29.8466 * {0.0724} 7.0237

-3736 -3767 32.1017 ** 27.4368 {0.0422} {0.1234} 6.7854 8.3749

-1643 26.3161 {0.1556} 7.9022

{1.0000} {0.2157} {0.9949} {0.9524} {0.6394} {0.9999} {0.1368} {0.2490} {0.9966} {0.9973} {0.9891} {0.9925} Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively. Q(20) and Q2(20) are Ljung-Box portmanteau test of white noise for linear and non-linear (squared) standardized residual.

104

There are also significant volatility responses to the Feds news. In eight out of twelve markets, significant responses are shown, and except for two markets, all show positive news coefficient suggesting a significantly higher volatility in response to the Feds news. This finding is consistent with the current literature, e.g. Wongswan (2009) and Bomfim (2003). It can be argued that, given the leading role of the US economy, the Asia Pacific market participants would closely monitor the USs policy announcements. Therefore, in approaching an impending Feds target rate announcement, even when the Asia Pacific market participants correctly guess the direction of the possible target rate change they were unsure about the magnitude of such an action. As a result, when there is a surprise in the announcement the homogeneity evaporates and instead would lead to heterogeneity of beliefs amongst the market participants. This then leads to a higher trading volume and eventually results in a higher volatility on the announcement day. However, in two markets (Hong Kong and Malaysia), the Feds news shows a volatility calming effect in the daily horizon. For Malaysia, there is a mixed response where the volatility was higher in the overnight horizon, but was lower in the intradaily horizon.42

42

For brevity, estimations results for the macroeconomic announcement dummies are omitted.

Instead, the main results are summarized here. The USs Macroeconomic announcement dummies show significant influence on the conditional volatility of Asia Pacific stock markets returns. In general, the USs GDP, Unemployment (UE) and Retail sales (RET) announcements tend to increase Asia Pacific markets volatility. Meanwhile, CPI announcements tend to reduce volatility in Asia Pacific stock markets in five out of twelve cases. In addition, the inclusion of the macro dummies does not significantly alter the interest rate news coefficients. All models were re-estimated without the 105

For the speed of adjustment, some evidence of tardiness in volatility response has been found. In four cases (Indonesia, New Zealand, the Philippines and Singapore), volatility response is significant only in intradaily horizon, and during both overnight and intradaily horizons in other two cases (Korea and Malaysia). Table 5.4 - The Feds target interest rate news spillover effects
This table reports the estimation results of the EGARCH models described in (5.2a) and (5.2b) for Asia Pacific stock markets daily returns against the overall surprises of the Fed. H0 is daily horizon over the close on calendar day t-1 to the close on calendar day t; H1 is overnight horizon over the close of day t-1 to the open on day t; and H2 is intraday horizon measured over the open on day t to the close on day t. P-values are in braces.

y t = [RHS of (5.1a)] + News News t 1

(5.2a) (5.2b)

ln ht = [RHS of (5.1b)] + News Newst 1 + MacroAnn,i MacroAnni ,t 1

i

Australia China Hongkong Indonesia Japan Korea Malaysia New Zealand Philippines Singapore Taiwan Thailand

H0 -0.1745 ** {0.0413} -0.4491 *** {0.0036} 0.5395 {0.2629} -0.1590 {0.9026} -0.3141 ** {0.0157} -0.0651 {0.2768} -0.2030 ** {0.0490} -0.1598 {0.6918} 0.1713 {0.8014} -0.3190 {0.4699} -0.4076 * {0.0552} -0.0828 ** {0.0106}

The Feds' target interest rate news spillover effects FedNews FedNews H1 H2 H0 H1 -0.1018 0.0112 0.0549 -0.0320 {0.6925} {0.8601} {0.7617} {0.8655} -0.3315 * -0.4178 ** 0.4271 ** 0.4293 ** {0.0661} {0.0216} {0.0212} {0.0346} -0.7201 * 0.5448 -0.2305 ** -0.1731 {0.0528} {0.2423} {0.0475} {0.3559} 1.5096 -0.8898 1.1608 0.6583 {0.2122} {0.2290} {0.5380} {0.7025} -0.1471 -0.2895 0.7834 1.1273 {0.5456} {0.8864} {0.7306} {0.1352} -0.0031 -0.0043 2.7768 ** 2.7770 * {0.9005} {0.7487} {0.0182} {0.0832} -0.3826 ** -0.3015 *** -0.6926 *** -1.0084 *** {0.0365} {0.0000} {0.0000} {0.0000} -0.3660 -0.0022 0.6645 *** 0.5507 {0.3833} {0.9615} {0.0018} {0.2143} 0.4646 -0.0593 0.4078 ** 0.6727 {0.3969} {0.7984} {0.0212} {0.1641} -0.8209 ** 0.1699 -0.0051 -0.2001 {0.0181} {0.4997} {0.9834} {0.4596} -0.9175 -0.9194 ** 0.0800 0.0275 {0.2626} {0.0127} {0.5518} {0.8837} -0.0811 ** -0.0812 * -0.2599 0.3290 {0.0141} {0.0836} {0.5914} {0.7446}

H2 -0.2721 {0.6435} 0.2423 {0.7922} -0.2326 {0.2880} 3.3660 ** {0.0253} 0.1770 {0.8716} 2.7767 ** {0.0407} 2.2832 ** {0.0284} 0.0146 ** {0.0173} 1.7377 * {0.0766} 0.6754 *** {0.0012} -0.2459 {0.3544} -0.0824 {0.7262}

macro announcement dummies and the resulting estimates for the interest rate news coefficients are qualitatively the same.

106

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

5.3.3. The ECBs target interest rate news spillover effects The quasi-maximum likelihood estimates of the EGARCH model for the spillover effects of the Feds target rate surprises as modeled in (5.2a) and (5.2b) are reported in Table 5.5. Table 5.5 - The ECBs target interest rate news spillover effects
This table reports the estimation results of the EGARCH models described in (5.2a) and (5.2b) for Asia Pacific stock markets daily returns against the overall surprises of the ECB. H0 is daily horizon over the close on calendar day t-1 to the close on calendar day t; H1 is overnight horizon over the close of day t-1 to the open on day t; and H2 is intraday horizon measured over the open on day t to the close on day t. P-values are in braces.

y t = [RHS of (5.1a)] + News News t 1

(5.2a) (5.2b)

ln ht = [RHS of (5.1b)] + News News t 1 + MacroAnn,i MacroAnni ,t 1

i

Australia China Hongkong Indonesia Japan Korea Malaysia New Zealand Philippines Singapore Taiwan Thailand

H0 -2.1945 * {0.0673} -1.8207 {0.2300} -2.6935 {0.5553} -2.6802 {0.2069} -0.1910 ** {0.0440} -3.7367 {0.4744} -2.0355 *** {0.0000} -2.2772 {0.3952} -0.2813 ** {0.0142} -0.3190 {0.4699} -3.3250 {0.4022} -2.9932 {0.3112}

The ECB' target interest rate news spillover effects ECBNews ECBNews H1 H2 H0 H1 -0.8273 -0.5315 *** 2.6766 0.2084 {0.6622} {0.0000} {0.1581} {0.9108} -1.8471 -1.3179 * -4.4646 ** -5.3809 *** {0.3802} {0.0741} {0.0283} {0.0078} -0.3287 *** -2.0008 1.1340 0.1320 {0.0000} {0.2333} {0.3654} {0.9107} -1.8256 ** -3.8748 *** -7.5952 -7.5083 {0.0494} {0.0000} {0.8727} {0.6584} 0.9374 -0.2346 *** 2.1867 7.2784 ** {0.1356} {0.0000} {0.5753} {0.0459} -4.4567 ** -3.7688 -1.7551 -4.5725 {0.0377} {0.4532} {0.1991} {0.7678} -2.3611 -0.1056 -0.6804 -3.2632 {0.1955} {0.6583} {0.7925} {0.2638} -0.3030 -3.4020 *** 3.6224 * 2.7669 {0.9110} {0.0000} {0.0815} {0.1839} -0.1393 *** -1.1326 *** 9.8749 ** 4.0324 ** {0.0000} {0.0006} {0.0142} {0.0236} -0.8209 ** 0.1699 -0.0051 -0.2001 {0.0181} {0.4997} {0.9834} {0.4596} -1.3015 -2.2022 *** 1.1363 0.0115 {0.7568} {0.0000} {0.4545} {0.9955} -3.3298 -1.4382 *** -1.8600 0.5023 {0.4280} {0.0000} {0.4071} {0.9040}

H2 1.7222 {0.5071} -8.4784 * {0.0955} 2.1926 {0.1108} 7.3878 *** {0.0000} 0.3567 {0.9352} -2.7765 {0.2274} 1.0452 ** {0.0366} 2.8093 * {0.0980} 8.7012 *** {0.0000} -0.6754 *** {0.0012} -1.0157 {0.5102} 16.0166 {0.1439}

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

107

Similar to the Feds new effects, the ECBs news had a negative spillover influence on the conditional means of all twelve stock returns over at least one horizon. It can be conjectured that given the increasing real integration between the Asia Pacific region and the euro area an unexpected rise in the ECBs main refinancing rate would have a negative impact on the euro area production leading to an expectation of lower demand for export goods from the Asia Pacific region. This could ultimately prove to be a negative influence for the Asia Pacific stock markets. In general, the ECB news was absorbed slowly as in six cases where the news coefficient is significant in the intradaily horizon and not in the overnight horizon. In three markets, however, the reverse is the case. This is in contrast with what has been found for the Feds news where the results of the speed of adjustment are mixed across the markets. It appears that the market participants in the Asia Pacific region were less than clear on the information content of the ECBs news, specifically on its additional contribution to tradable information on the heels of the Feds interest rate movements. Strong evidence for a volatility raising influence of the ECBs target rate news has also been documented. Five out of twelve markets show a significantly higher volatility in response to the ECB news. On the other hand, China and Singapore show a volatility reducing influence. Considering the volatility stimulating effect of the Feds news on these two markets as reported above, the ECBs market calming effect suggests that the ECBs news helps to resolves the additional uncertainty that was created by the earlier arrival of the Feds news in these two markets. There is also some evidence of the volatility response being either concentrated or persisting into the intradaily horizon. In four markets (Indonesia,
108

Malaysia, New Zealand and Singapore), the significant volatility response is shown only in the intradaily horizon and one country (Korea) showing only in the overnight horizon. Furthermore, two markets (China and the Philippines) reveal higher volatility in the overnight horizon that is persisting into the intradaily horizon.43 5.4. Robustness checks 5.4.1. Joint spillover effect of the target interest rate news The joint spillover effects of the Feds and the ECBs target rate news has been investigated as modeled in equations (5.3a) and (5.3b) and the results are reported in Table 5.6. Overall, the results suggest that the separate effects of the Feds and the ECBs news as reported in the previous sections are robust. In general, while the Feds news effect (on both moments) and the ECBs news effect on the conditional mean are marginally weaker, the ECBs news effect on the conditional volatility is much stronger with five cases showing lower volatility. In addition to China and Singapore as reported above, the ECBs market calming effect has also been shown in stock markets of Indonesia and Korea (overnight only), and Malaysia (intraday only). This finding suggests that the ECBs news provided further

43

The euro areas macroeconomic announcement dummies tend to have an opposite impact on the

volatility of the Asia Pacific stock markets with the USs ones. The euro areas GDP, Unemployment (UE) and Current account balance (CAB) announcements tend to reduce the Asia Pacific stock markets volatility while the reverse is observed for CPI and External trade (ET). The Retail trade growth rate announcements show unclear impact on the volatility with even cases of four increases and decreases.

109

information that helps to resolve heterogeneous beliefs caused by the earlier arrival of the Feds news. Table 5.6 - The US Feds and the ECBs Joint spillover effects
This table reports the estimation results of EGARCH models described in (5.3a) and (5.3b) for Asia Pacific stock markets daily returns against the Feds and the ECBs news. H0 is daily horizon over the close on calendar day t-1 to the close on calendar day t; H1 is overnight horizon over the close of day t-1 to the open on day t; and H2 is intraday horizon measured over the open on day t to the close on day t. P-values are in braces.

y t = [RHS of (5.1a)] + FedNews FedNews t 1 + ECBNews ECBNews t 1

(5.3a) (5.3b)

ln ht = [RHS of (5.1b)] + FedNews FedNewst 1 + USMacroAnn,iUSMacroAnni ,t 1

i

+ ECBNews ECBNewst 1 + EuroareaMacroAnn,i EuroareaMacroAnni ,t 1

i

Panel A Conditional mean equation

Conditional mean equation H0 -0.1232 {0.4961} -0.5441 ** {0.0112} 0.6150 {0.1881} -0.2089 {0.8750} -0.0611 {0.4584} -0.1463 {0.8751} -0.1589 {0.3746} -0.5335 {0.5293} -0.1060 {0.3299} -0.4212 {0.1575} -0.8217 ** {0.0451} -0.1932 {0.6641} FedNews H1 -0.1079 *** {0.0000} -0.4180 {0.2562} -0.5520 *** {0.0014} 1.3655 {0.3687} -0.4797 {0.5720} -1.2331 {0.3341} -0.3695 {0.1381} -0.5449 {0.1184} -0.0991 {0.8464} -1.2050 ** {0.0159} -1.1363 {0.1200} -0.1023 {0.8369} H2 0.0172 {0.5221} -0.9594 *** {0.0057} 0.6353 {0.1082} -0.6528 {0.3662} -0.1120 *** {0.0088} -0.1217 {0.8894} -0.0964 {0.6229} -0.6298 {0.2641} -0.0964 {0.5388} 0.3874 {0.1086} -0.9822 *** {0.0021} 0.1110 {0.7759} H0 -2.4434 {0.1559} -0.3621 {0.4394} -1.4131 {0.4840} -3.6856 *** {0.0003} -0.1769 *** {0.0000} -2.0348 {0.7047} -2.1111 {0.3290} 0.0839 {0.2123} -0.9632 *** {0.0000} -4.7732 ** {0.0206} -2.4362 {0.8595} 3.2088 {0.3429} ECBNews H1 -1.0608 {0.5968} -1.7944 {0.3286} 0.0553 {0.9941} -2.5274 {0.6220} 0.1476 {0.1380} -4.2634 {0.1063} -2.2535 {0.3209} 0.2627 {0.9284} -0.4072 {0.8702} -4.3616 * {0.0687} -1.0057 {0.8186} 2.8950 {0.4883} H2 -0.6397 {0.1956} -0.7935 *** {0.0071} -1.5302 {0.4525} -0.7014 {0.5371} -0.2941 *** {0.0003} -2.1445 {0.6198} -0.1356 {0.7009} 0.2969 {0.9398} -0.9475 * {0.0519} 0.4173 {0.3776} -2.5276 {0.2628} 0.3418 {0.3014}

Australia China Hongkong Indonesia Japan Korea Malaysia New Zealand Philippines Singapore Taiwan Thailand

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

110

Table 5.6 - Continued

Panel B. Conditional volatility equation
Conditional variance equation H0 -0.1610 {0.4272} 1.4542 *** {0.0000} -0.4011 {0.1916} 1.0634 {0.3595} 1.0114 {0.1933} -0.4939 {0.4039} -0.9978 *** {0.0000} 0.1584 {0.9085} 0.6913 ** {0.0417} 0.2237 {0.4425} -0.2630 {0.3678} 0.3154 {0.1895} FedNews H1 -0.0771 {0.7237} 0.6417 ** {0.0129} -0.0686 {0.7446} 1.5114 {0.4824} 1.0046 {0.1071} -0.2800 {0.3153} -1.3450 ** {0.0268} 0.1122 {0.5580} 0.5517 {0.2589} 0.1366 {0.3534} 0.0060 {0.9721} 0.1165 {0.6752} H2 -0.7153 {0.2175} 1.8971 ** {0.0138} -0.3970 *** {0.0022} 5.0306 *** {0.0042} 0.2962 ** {0.0475} -0.4797 {0.1463} 0.7659 {0.2079} -0.3292 {0.3490} 0.7117 ** {0.0253} 0.5970 *** {0.0002} -0.2395 {0.2664} -0.7192 {0.3441} H0 2.5155 ** {0.0151} -6.3297 *** {0.0057} 1.6597 {0.2009} 2.4692 {0.8383} 8.3465 ** {0.0411} -1.6817 {0.2031} 1.5115 {0.1427} 2.0956 *** {0.0000} 1.4917 *** {0.0000} -4.9674 *** {0.0064} -0.8588 {0.4233} -1.8230 {0.5634} ECBNews H1 -0.4095 {0.8040} -7.2225 *** {0.0000} 1.8742 {0.2915} -4.7839 * {0.0941} 6.7227 * {0.0573} -4.0809 ** {0.0215} -0.3509 {0.9017} 3.0473 *** {0.0073} 4.2443 ** {0.0196} -5.5828 *** {0.0060} -0.1335 {0.9325} 0.8077 {0.7177} H2 5.3135 {0.1280} -5.1569 *** {0.0000} -1.7758 {0.1997} 4.5096 {0.4189} 1.6115 * {0.0839} -1.4904 {0.5218} -1.8658 *** {0.0000} 2.4559 *** {0.0000} 2.5414 *** {0.0003} -11.9295 *** {0.0000} -0.9898 {0.5784} 4.6496 ** {0.0466}

Australia China Hongkong Indonesia Japan Korea Malaysia New Zealand Philippines Singapore Taiwan Thailand

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

5.4.2. The role of indirect impacts Figure 5.2 - Direct and indirect effects
The Feds/ECBs news The US/euro areas stock markets
(3)

(1)

(2)

The AsiaPacific stock markets

As the effect of the US Feds and the ECBs news could be direct (i.e. market participants in Asia Pacific countries react to the announcements, direction (3) of Figure 5.2) or indirect (i.e. Asia Pacific markets react to changes in the US and the EU markets, which followed the announcements, directions (1) and (2) of Figure 5.2). It is possible that the direct effect might be insignificant, but appears significant because the indirect effect is not properly controlled for.
111

We re-estimate our models with both the direct impact (i.e. the Feds or the ECBs target interest rate news effects) and the indirect impact (US or Euro area stock index returns spillovers to the Asia Pacific region) in the conditional mean and variance equations of the Asia Pacific stock market returns. The results are summarized in Table 5.7 where the interest rate news coefficients from the expanded models are presented. In general, the qualitative results shown are not significantly altered from what are reported in the previous sections. However, in the case of the Fed news effects on the first moment of the stock returns, fewer stock markets responded to the news in the first two horizons but more in the third horizon. This is not surprising due to the potential of the muticollinearity presents in the expanded model. We then establish the indirect impacts of the Feds and the ECBs target rate news as represented in directions (1) and (2) of Figure 5.2 via two-step estimations. First, Table 5.8 shows evidence for the Feds interest rate news effect on the daily S&P500 returns and the ECBs news effects on the DAX and the CAC. There is a significant negative impact on the mean, and a volatility raising effect on the volatility is present. This finding is consistent with that of Bernanke and Kuttner (2005) for the US stock markets and Bohl et al. (2008) for European stock markets. Table 5.7 - The Feds and the ECBs interest rate news effects estimated with indirect effects via index return spillovers
This table reports the estimation results of the EGARCH models below for Asia Pacific stock markets daily returns against the overall surprises of the Fed and the ECB. IndexReturns is the returns of S&P500 for the Fed and DAX and CAC for the ECB estimations. H0 is daily horizon over the close on calendar day t-1 to the close on calendar day t; H1 is overnight horizon over the close of day t-1 to the open on day t; and H2 is intraday horizon measured over the open on day t to the close on day t. P-values are in braces.

y t = [RHS of (5.1a)] + News News t 1 + Index Re turn t 1

(5.4a) (5.4b)

ln ht = [RHS of (5.1b)] + News News t 1 + MacroAnn,i MacroAnni ,t 1

i

+ Index Re turns

2 t 1

112

Table 5.7 - Continued

Panel A The Feds indirect effects
The Feds' target interest rate news spillover effects FedNews FedNews H1 H2 H0 H1 0.0141 -0.0411 0.2517 0.2769 {0.9575} {0.2398} {0.3144} {0.3989} -0.4553 ** -0.4555 ** 0.4815 * 0.5327 *** {0.0200} {0.0427} {0.0666} {0.0034} -0.2608 0.3628 *** -0.2080 ** -0.0212 {0.4647} {0.0004} {0.0214} {0.9189} 1.6597 -1.1716 * 1.7637 0.6627 {0.1661} {0.0669} {0.2338} {0.7169} 0.1614 -0.2934 -0.2669 -0.0093 {0.6679} {0.2134} {0.1835} {0.9774} 0.0091 -0.0111 *** 0.1509 -1.6040 *** {0.9017} {0.0000} {0.4491} {0.0000} -0.3243 -0.2744 * -0.7493 *** -1.0280 *** {0.4148} {0.0791} {0.0002} {0.0000} -0.3750 -0.0194 *** 0.6954 0.5459 {0.4861} {0.0000} {0.1972} {0.1792} 0.2999 -0.3488 *** 1.3921 *** 0.6163 {0.4145} {0.0000} {0.0002} {0.2164} -0.5339 ** -0.0249 0.1293 0.0287 {0.0141} {0.9464} {0.6581} {0.9181} -0.3238 0.3458 0.1020 0.0433 {0.7218} {0.1800} {0.5999} {0.8321} 0.1106 -0.0515 0.4978 0.4638 {0.8632} {0.7663} {0.1135} {0.2438}

Australia China Hongkong Indonesia Japan Korea Malaysia New Zealand Philippines Singapore Taiwan Thailand

H0 -0.1506 {0.5299} -0.4908 *** {0.0090} 0.1718 {0.5097} -0.0285 {0.9825} -0.2096 {0.4567} -0.2998 ** {0.0216} -0.1331 *** {0.0000} -0.1652 {0.7600} 0.7296 {0.6123} -0.2002 {0.6412} -0.3044 {0.6711} -0.0959 {0.9110}

H2 -1.0585 ** {0.0117} -0.0768 {0.9619} -0.0878 {0.6512} 5.9645 ** {0.0248} 0.0988 {0.7677} -1.4927 ** {0.0180} 0.9631 *** {0.0052} -1.8992 ** {0.0196} 2.1284 *** {0.0000} 0.7680 ** {0.0260} -0.6108 ** {0.0331} -0.4702 {0.3754}

Panel B The ECBs indirect effects

The ECB' target interest rate news spillover effects ECBNews ECBNews H1 H2 H0 H1 -2.0631 -0.3631 * 2.0195 ** 0.1062 {0.2485} {0.0648} {0.0419} {0.9549} -1.9630 -1.1685 *** -4.4415 ** -5.5143 *** {0.3300} {0.0000} {0.0371} {0.0028} -3.3740 0.8140 13.2268 *** -1.2278 {0.3051} {0.5120} {0.0035} {0.3596} -2.0260 -1.4663 -10.5022 -17.2596 {0.6400} {0.5645} {0.7472} {0.7293} 3.0025 -0.2777 9.4661 * 6.7898 ** {0.5167} {0.7696} {0.0966} {0.0268} 9.5177 * -3.9695 -3.5350 * -4.1679 * {0.0662} {0.4813} {0.0992} {0.0589} 1.5066 -0.0012 -1.7834 -3.4718 * {0.3793} {0.9974} {0.3539} {0.0555} 0.2286 -0.0164 3.7635 *** 2.7969 * {0.9404} {0.9217} {0.0011} {0.0537} -0.1212 1.6963 *** 9.7032 *** 4.5370 * {0.9431} {0.0000} {0.0000} {0.0583} 2.9796 ** 1.0243 -4.3639 *** -5.3413 *** {0.0131} {0.3010} {0.0048} {0.0065} -3.3632 -1.1684 0.5462 -0.8207 {0.4672} {0.4892} {0.8013} {0.6043} 1.0368 -2.4437 -1.9279 -0.5382 {0.7766} {0.4413} {0.4672} {0.6749}

Australia China Hongkong Indonesia Japan Korea Malaysia New Zealand Philippines Singapore Taiwan Thailand

H0 -3.3146 * {0.0681} -1.4039 {0.2170} 3.0495 {0.3143} -1.9644 {0.5903} 0.0006 *** {0.0000} -4.2448 * {0.0870} 2.1893 *** {0.0000} 2.2466 {0.4269} 6.6613 {0.1007} 3.1243 ** {0.0458} -4.8409 *** {0.0000} 1.8476 {0.5920}

H2 1.9815 *** {0.0001} -9.7233 *** {0.0063} -3.0868 * {0.0867} 7.9586 ** {0.0453} -0.2608 {0.9286} -2.7195 {0.1946} -11.5247 *** {0.0003} 11.0576 *** {0.0000} -26.9229 *** {0.0000} -9.9260 *** {0.0000} -3.3095 * {0.0681} 14.6425 {0.2813}

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

113

Table 5.8 - The Feds and the ECBs interest rate news effects on domestic markets
This table reports the estimation results of the EGARCH models below for the US and the euro areas stock markets daily returns against the overall surprises of the Fed and the. ECB. y is the daily return of S&P500 for the Fed and DAX and CAC for the ECB estimations. P-values are in braces.

y t = [RHS of (5.1a)] + News News t 1

ln ht = [RHS of (5.1b)] + News News t 1 + MacroAnn,i MacroAnni ,t 1

i

(5.5a) (5.5b)

SP500 DAX CAC *** *** FedNews -1.3299 ECBNews -0.1124 -0.9322 {0.0000} {0.0000} {0.0123} FedNews 0.2019 * ECBNews 1.8955 ** -0.6489 {0.0698} {0.0198} {0.5720} Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

**

In addition, the literature shows significant spillover effects of the US markets to the Asia Pacific markets. Table 5.9 summarizes the spillover effects of the USs (S&P500) and the euro area stock markets (DAX and CAC) returns on the Asia Pacific stock markets. The Asia Pacific stock market returns are, in general, moved in the same direction as the overnight developments in the US and the euro area markets. The Asia Pacific markets returns and volatilities are higher in response to the overnight rally in the major markets. Thus, including both the Fed/ECB interest rate news and US/euro area stock index returns in our investigation would lead to multicollinearity issues. Consequently, the reduction of significance of the news coefficients as reported in Tables 5.7 is expected, as they are not precisely estimated. In order to improve the precision of the news coefficient estimations and to concentrate on the direct impact of the news, this study decided not to control for the indirect news effects. However, given the discussion of the potential multicollinearity between the direct and indirect effects, it appears to be difficult to distinguish the two effects. Therefore, there might be arguments that the results in this chapter seem to be expected given what has been documented in the literature. To tackle this, ChiSQ tests have been conducted to show the marginal contribution of
114

direct effect given indirect effect. The results shown in Table 5.10 reject the null hypothesis that the news coefficients are zero in the conditional mean and variance equations. Table 5.9 - Spillover effects of the USs and the euro areas stock markets on the Asia Pacific markets
This table reports the estimation results of the EGARCH models below for the Asia Pacific stock markets daily returns against the US and euro area stock markets. IndexReturns is the returns of S&P500 for the Fed and DAX and CAC for the ECB estimations. P-values are in braces.

y t = [RHS of (5.1a)] + Index Re turn t 1

ln ht = [RHS of (5.1b)] + MacroAnn,i MacroAnni ,t 1 + Index Re turns t21

i

(5.6a) (5.6b)

Panel A Spillover effects of the USs stock markets on the Asia Pacific markets
H0 0.0400 *** {0.0000} -0.0005 {0.5547} 0.1143 *** {0.0000} -0.0238 {0.4340} 0.1205 *** {0.0000} 0.0019 *** {0.0000} -0.0267 *** {0.0000} 0.0005 {0.9577} 0.0696 *** {0.0009} 0.1233 *** {0.0000} 0.1190 *** {0.0001} 0.0781 *** {0.0000} USIndex H1 0.1447 *** {0.0000} -0.0212 {0.3373} 0.5815 *** {0.0000} -0.0283 ** {0.0394} 0.4477 *** {0.0000} 0.0019 *** {0.0000} 0.0606 *** {0.0000} -0.0036 {0.6990} 0.0660 *** {0.0010} 0.4833 *** {0.0000} 0.6083 *** {0.0000} 0.2001 *** {0.0000} H2 -0.0680 {0.0000} 0.0187 *** {0.0002} 0.4702 *** {0.0000} -0.0049 {0.4621} 0.2516 *** {0.0000} 0.0008 {0.7074} -0.0691 *** {0.0000} 0.0023 *** {0.0000} 0.0644 *** {0.0000} 0.4239 *** {0.0000} 0.5174 *** {0.0000} 0.1329 *** {0.0000} H0 0.0107 {0.1453} 0.0069 ** {0.0264} 0.0025 {0.2686} 0.0124 {0.2177} 0.0157 *** {0.0000} 0.0471 *** {0.0000} 0.0035 {0.4894} -0.0007 {0.6888} 0.0085 {0.1467} 0.0246 *** {0.0026} 0.0099 *** {0.0020} -0.0006 {0.9001} USIndex H1 0.0145 {0.1482} 0.0041 {0.2561} 0.0057 {0.1254} 0.0054 {0.2046} 0.0234 ** {0.0166} 0.0469 ** {0.0379} 0.0010 {0.8032} 0.0004 {0.7959} 0.0076 {0.1029} 0.0264 *** {0.0000} 0.0114 *** {0.0037} -0.0028 {0.5188} H2 -0.2721 {0.6435} 0.0821 *** {0.0000} 0.0274 * {0.0674} 0.0279 *** {0.0000} -0.0005 {0.7501} 0.0071 {0.1384} 0.1287 ** {0.0500} -0.0089 {0.2801} 0.0488 ** {0.0344} 0.1011 *** {0.0000} 0.0419 *** {0.0000} 0.0195 ** {0.0145}

Australia China Hongkong Indonesia Japan Korea Malaysia New Zealand Philippines Singapore Taiwan Thailand

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

115

Table 5.9 Continued

Panel B Spillover effects of the euro areas stock markets on the Asia Pacific markets
Mean equation H0 0.2089 ** {0.0224} 0.4299 *** {0.0001} 0.3801 ** {0.0477} 0.2079 {0.4561} 0.2950 {0.1060} 0.5747 *** {0.0089} 0.0756 {0.5019} -0.0831 *** {0.0000} 0.3485 ** {0.0354} 0.0994 {0.4294} 0.1304 {0.5115} -0.0762 {0.6499} DAX H1 0.1713 ** {0.0290} 0.2816 {0.1402} 0.2987 * {0.0842} 0.1475 *** {0.0000} 0.0776 {0.7534} 0.2505 {0.3766} 0.0258 {0.8346} 0.0183 {0.8419} 0.3801 ** {0.0355} 0.4096 *** {0.0000} 0.1891 {0.4603} 0.0889 *** {0.0029} DAX H1 0.0189 {0.4165} -0.0035 {0.9355} 0.0178 {0.2736} 1.8972 {0.1694} 2.5416 *** {0.0000} 0.1025 {0.4709} 0.0033 {0.8986} 0.0118 {0.9132} 0.0719 {0.2850} -0.0499 {0.1397} -0.0583 {0.1574} 0.0640 ** {0.0426} H2 H0 0.0030 0.1155 *** {0.8728} {0.0000} 0.0007 -0.0054 *** {0.9819} {0.0000} 0.2151 *** 0.2631 *** {0.0024} {0.0000} 0.1322 *** 0.0446 * {0.0051} {0.0864} -0.0389 0.2515 *** {0.6301} {0.0000} 0.5700 ** 0.0376 * {0.0168} {0.0887} 0.0199 ** 0.0636 *** {0.0144} {0.0000} -0.0721 *** 0.0113 *** {0.0000} {0.0000} 0.1195 0.0479 *** {0.1456} {0.0057} 0.2293 *** 0.2593 *** {0.0008} {0.0000} 0.0469 0.1866 *** {0.6443} {0.0000} 0.1241 *** 0.2012 *** {0.0007} {0.0000} Volatility equation H2 -0.0198 {0.5856} -0.0898 {0.4893} 0.0438 {0.1314} 1.4310 ** {0.0281} -0.0305 {0.4528} -0.0270 {0.9409} 0.0710 {0.6435} 2.0757 *** {0.0000} 1.4485 ** {0.0105} -0.0510 {0.2195} 0.0241 * {0.0748} 0.0899 {0.4139} H0 0.0486 *** {0.0000} 0.0026 * {0.0791} 0.0047 * {0.0598} 0.0214 * {0.0593} 0.0082 *** {0.0000} 0.0028 * {0.0799} 0.6804 {0.7925} 0.0024 {0.1103} 0.0112 *** {0.0060} 0.0753 *** {0.0000} 0.0320 *** {0.0044} 0.0010 {0.2413} CAC H1 0.1645 *** {0.0000} -0.0190 {0.2448} 0.4494 *** {0.0000} 0.0505 *** {0.0000} 0.3987 *** {0.0000} -0.0071 {0.8032} 0.0885 *** {0.0000} 0.0104 {0.1432} 0.0378 ** {0.0162} 0.3751 *** {0.0000} 0.3671 *** {0.0000} 0.2366 *** {0.0000} CAC H1 0.0394 *** {0.0005} 0.0015 {0.9190} 0.0309 *** {0.0057} 0.0197 * {0.0748} 0.0197 {0.2710} 0.0028 * {0.0537} 0.0313 *** {0.0009} 0.0011 {0.2832} 0.0031 {0.8176} 0.0466 *** {0.0000} 0.0469 *** {0.0002} 0.0281 ** {0.0338} H2 0.0357 *** {0.0000} 0.0007 {0.8811} 0.2245 *** {0.0000} -0.0045 {0.6027} 0.1462 *** {0.0000} 0.0375 {0.3874} 0.0306 *** {0.0000} 0.0002 {0.5054} 0.0467 *** {0.0000} 0.1407 *** {0.0000} 0.2148 *** {0.0000} 0.0425 *** {0.0000}

Australia China Hongkong Indonesia Japan Korea Malaysia New Zealand Philippines Singapore Taiwan Thailand

Australia China Hongkong Indonesia Japan Korea Malaysia New Zealand Philippines Singapore Taiwan Thailand

H0 0.0056 {0.7909} 0.0305 {0.8681} 0.0123 {0.8988} 2.2159 {0.2328} 0.1225 * {0.0939} -0.0174 {0.9065} 2.0355 *** {0.0000} 0.0343 {0.7924} 1.2731 *** {0.0060} -0.0206 {0.4576} -0.0415 {0.2089} 0.2643 *** {0.0035}

H2 0.0144 {0.4154} 0.0062 {0.8820} 0.0546 *** {0.0009} 0.0059 {0.4851} -0.0407 {0.1140} 0.0028 {0.4766} 0.0885 * {0.0898} 0.0242 *** {0.0000} -0.0236 {0.4744} 0.1087 *** {0.0000} 0.1106 *** {0.0000} 0.0497 {0.1608}

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

116

Table 5.10 - Marginal contribution of direct effect given indirect effect

This table reports the ChiSQ test of the null hypothesis that the News coefficients are zero in the EGARCH models below for Asia Pacific stock markets daily returns against the overall surprises of the Fed and the ECB. IndexReturns is the returns of S&P500 for the Fed and DAX for the ECB estimations. H0 is daily horizon over the close on calendar day t-1 to the close on calendar day t; H1 is overnight horizon over the close of day t-1 to the open on day t; and H2 is intraday horizon measured over the open on day t to the close on day t. P-values are in braces.

y t = [RHS of (5.1a)] + News News t 1 + Index Re turn t 1

2 t 1 i

ln ht = [RHS of (5.1b)] + News Newst 1 + MacroAnn,i MacroAnni ,t 1 + Index Re turns

(5.7a) (5.7b)

Panel A Marginal contribution of the Feds direct effect given indirect effect

Test if the Fed news coefficents to be zero in the mean and variance equations Mean equation Variance equation Both mean and variance equations H0 H1 H2 H0 H1 H2 H0 H1 H2 Australia 0.3911 0.0044 1.6083 1.0021 0.9908 3.2501 *** 1.9950 0.8582 7.5015 *** {0.5317} {0.9472} {0.2047} {0.3168} {0.3195} {0.0000} {0.1578} {0.3542} {0.0000} *** ** *** *** *** ** ** China 2.6732 5.6056 4.7154 1.3456 8.7563 0.5976 4.6572 1.2022 5.6614 *** {0.0000} {0.0200} {0.0000} {0.0000} {0.0031} {0.4395} {0.0309} {0.0273} {0.0000} Hongkong 0.4299 -0.2608 7.3039 *** 5.2919 ** -0.0212 0.0343 2.5950 0.0984 1.3971 ** {0.5121} {0.4647} {0.0000} {0.0214} {0.9189} {0.8531} {0.1072} {0.7538} {0.0237} Indonesia 0.0002 1.2458 3.3575 * 3.9220 ** 0.1609 5.0389 ** 0.8811 0.4439 5.8698 *** {0.9879} {0.2644} {0.0669} {0.0477} {0.6884} {0.0248} {0.3479} {0.5052} {0.0000} Japan 0.1081 0.0459 1.4256 1.1654 0.0007 0.0447 0.0085 0.0521 0.5978 {0.7423} {0.8303} {0.2325} {0.2803} {0.9786} {0.8325} {0.9265} {0.8194} {0.4394} Korea 4.9530 ** 0.0153 4.4166 *** 0.3001 6.3934 *** 5.5970 ** 8.0760 *** 2.7869 *** 5.5150 ** Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

117

Table 5.10 Continued

Test if the Fed news coefficents to be zero in the mean and variance equations Mean equation Variance equation Both mean and variance equations H0 H1 H2 H0 H1 H2 H0 H1 H2 {0.0260} {0.9017} {0.0000} {0.5838} {0.0000} {0.0180} {0.0000} {0.0000} {0.0189} Malaysia 0.2404 ** 0.6644 4.8559 ** 2.8756 *** 4.6555 *** 1.3855 ** 4.0119 ** 2.1133 2.2302 {0.0239} {0.4150} {0.0276} {0.0000} {0.0000} {0.0239} {0.0452} {0.1460} {0.1353} New Zealand 0.0933 0.4852 5.8499 *** 1.6632 1.8040 5.4501 ** 2.7212 * 1.6154 5.3331 ** {0.7600} {0.4861} {0.0000} {0.1972} {0.1792} {0.0196} {0.0990} {0.2037} {0.0209} Philippines 0.2568 0.2792 0.2974 *** 4.1451 *** 1.0141 1.5672 *** 0.1756 0.1552 4.5515 *** {0.6123} {0.5972} {0.0050} {0.0002} {0.3139} {0.0007} {0.6752} {0.6936} {0.0001} Singapore 0.6270 1.9742 ** 0.0059 0.1542 0.0287 5.4829 ** 0.5393 0.1776 1.6588 {0.4285} {0.0160} {0.9388} {0.6946} {0.9181} {0.0192} {0.4627} {0.6734} {0.1978} Taiwan 1.1616 0.1257 2.2832 0.1572 0.0449 3.2086 * 1.0465 0.1572 3.8822 ** {0.2811} {0.7230} {0.1308} {0.6917} {0.8322} {0.0733} {0.3063} {0.6917} {0.0488} Thailand 0.6547 0.0297 0.2040 2.1310 1.4927 0.7856 0.4882 0.2356 0.4086 {0.4184} {0.8669} {0.6515} {0.1444} {0.2218} {0.6515} {0.4848} {0.6274} {0.5227} Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

118

Table 5.10 Continued

Panel B Marginal contribution of the ECBs direct effect given indirect effect

Test if the ECB news coefficents to be zero in the mean and variance equations Mean equation Variance equation Both mean and variance equations H0 H1 H2 H0 H1 H2 H0 H1 H2 * * ** *** *** Australia 3.3291 1.3315 3.4110 4.1405 0.0032 14.3748 8.0696 0.8669 14.6920 *** {0.0681} {0.2485} {0.0648} {0.0419} {0.9549} {0.0001} {0.0045} {0.3518} {0.0001} China 1.5239 0.9489 9.8611 *** 4.3473 ** 8.9218 *** 7.5472 *** 1.2730 1.1964 9.4701 *** {0.2170} {0.3300} {0.0000} {0.0371} {0.0028} {0.0060} {0.2592} {0.2740} {0.0021} Hongkong 1.0126 1.0517 0.4300 8.5290 *** 0.8391 2.9337 * 3.4213 * 0.4046 2.9241 * {0.3143} {0.3051} {0.5120} {0.0035} {0.3596} {0.0867} {0.0644} {0.5247} {0.0873} Indonesia 0.0259 0.0710 0.7205 0.2822 0.1111 6.6381 *** 0.3946 0.0970 3.5909 * {0.8722} {0.7898} {0.3960} {0.5953} {0.7389} {0.0100} {0.5299} {0.7554} {0.0581} Japan 8.1232 *** 0.4205 0.0858 2.7608 * 4.9059 ** 0.0080 2.7604 * 0.6264 ** 0.0187 {0.0000} {0.5167} {0.7696} {0.0966} {0.0268} {0.9286} {0.0966} {0.0429} {0.8912} Korea 4.4123 ** 3.1422 * 0.4252 5.8880 ** 5.2698 ** 1.8176 0.0021 0.9632 *** 0.0093 {0.0357} {0.0763} {0.5144} {0.0152} {0.0217} {0.1776} {0.9632} {0.0071} {0.9233} Malaysia 6.7167 *** 0.7728 0.0333 0.8594 3.6670 * 12.9656 *** 4.2681 ** 3.1231 * 12.7208 *** {0.0000} {0.3793} {0.8553} {0.3539} {0.0555} {0.0003} {0.0388} {0.0772} {0.0004} New Zealand 0.6313 0.0056 0.0097 10.5827 *** 3.7228 * 9.2866 *** 0.2993 0.6318 62.1416 *** {0.4269} {0.9404} {0.9217} {0.0011} {0.0537} {0.0000} {0.5843} {0.4267} {0.0000} Philippines 2.6948 0.0051 4.0465 *** 1.9829 *** 3.5862 * 5.2427 *** 0.9239 2.1534 4.4911 *** {0.1007} {0.9431} {0.0000} {0.0000} {0.0583} {0.0000} {0.3365} {0.1423} {0.0000} Singapore 3.9904 ** 6.1521 ** 1.0700 7.9483 *** 7.4145 *** 9.2804 *** 11.6104 *** 12.2145 *** 8.4470 *** {0.0458} {0.0131} {0.3010} {0.0048} {0.0065} {0.0000} {0.0007} {0.0005} {0.0000} Taiwan 6.3426 *** 0.5285 0.4782 0.0634 0.2686 3.3282 * 6.1709 ** 0.3256 0.7173 {0.0000} {0.4672} {0.4892} {0.8013} {0.6043} {0.0681} {0.0130} {0.5683} {0.3970} Thailand 0.2872 0.0805 0.5929 0.5285 0.1759 1.1606 0.7080 0.1595 1.0863 {0.5920} {0.7766} {0.4413} {0.4672} {0.6749} {0.2813} {0.4001} {0.6896} {0.2973} Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

119

5.4.3. Policy cycle impacts Evidence on cyclical effects of the Feds news has been controversially documented in the literature where Andersen et al. (2007) find no evidence of state dependence in the stock markets response to monetary news. However, using a more accurate measure of monetary news, a longer sample period, and multiple proxies for macroeconomic state, Basistha and Kurov (2008) find strong evidence of such state dependence for the Feds news. However, notwithstanding the growing importance of the euro area as well as the inflation targeting framework conducted by the ECB, which is claimed to be superior to other monetary policy regimes, there is no evidence on the cyclical impacts of the ECBs target rate news on the stock markets. An investigation to provide evidence of cyclical variation in the response of stocks to monetary news should be useful not only to the ECB policymakers by helping them understand and foresee the ultimate effect of a target rate change on the stock market so that they could manage their monetary policy more effectively during different business cycles; but also to market participants by helping them design and adjust trading strategies in response to such shocks in the ECBs target interest rate. It is possible that the relative role of the monetary policy news can depend on the policy stance at the time. For instance, during the tightening cycle, unexpected rate changes might have stronger news impact as lower than expected rate hikes might lead to relief rallies in the stock markets. Over our sample, the monetary
120

policy stance of the Fed and the ECB includes both contraction and expansion cycles. For the Fed, the expansion cycle was from Jan 1999 to May 2004, and the tightening phase was from Jun 2004 to Dec 2006. For the ECB, these were May 2001 to Nov 2005 and Jan 1999 to Apr 2001, respectively. Sub-sample analyses for each of the two policy cycles for both central banks were conducted and the results are reported in Table 5.11. For the Fed news spillovers, there is some evidence suggesting that the sensitivity to the news is larger during the contraction cycle. Although, the signs of the news coefficients are consistent between the two cycles, the magnitudes are larger, in general, for both the first and second moment responses during the tightening period. This suggests that the markets were more sensitive to unexpected movements in the Feds target rate during the period of rising interest rate cycle. However, there is no clear evidence for such sample specific market responses to the ECBs news. 5.4.4. Asymmetric effects The impacts of positive and negative interest rate news are estimated separately and the results are reported in Table 5.12. In four countries, the conditional volatilities respond mostly to unexpectedly higher rate announcements from the Fed; however, in other countries there are no consistent responses. Similarly, in the case of the ECB news, the patterns of asymmetric responses between the two types of news are not uniform across the markets.

121

Table 5.11 - The policy cycle impacts

This table reports the estimation results of the EGARCH models described in (5.2a) and (5.2b) for Asia Pacific stock markets daily returns against the overall surprises of the Fed and the ECB over the expansionary and contractionary periods. H0 is daily horizon over the close on calendar day t-1 to the close on calendar day t; H1 is overnight horizon over the close of day t-1 to the open on day t; and H2 is intraday horizon measured over the open on day t to the close on day t. P-values are in braces.

y t = [RHS of (5.1a)] + News News t 1

(5.2a) (5.2b)

ln ht = [RHS of (5.1b)] + News Newst 1 + MacroAnn,i MacroAnni ,t 1

i

Panel A The Feds policy cycle impacts

The Feds' target interest rate news spillover effects

Jan 1999 - May 2004 (expansion)

Australia

China

Hongkong

H0 -0.1676 {0.4225} -0.2125 {0.3091} -0.6727 *** {0.0000}

FedNews H1 -0.1098 {0.5795} -0.3213 {0.2899} -1.4076 {0.2985} H2 0.0345 {0.3889} -0.2573 *** {0.0000} 0.6771 {0.1850} H0 0.0340 {0.8781} 0.3045 {0.5791} -0.0982 {0.5519} H2 -0.2192 {0.6222} 0.0309 {0.8755} -0.0925 {0.7065}

FedNews H1 0.0087 {0.9611} 0.4275 {0.3504} -0.0933 {0.7364}

Indonesia

Japan 0.2052 {0.2292} -0.9928 0.2153 {0.3948} -0.2035 0.8864 {0.2466} -0.1915

Korea

-0.5139 {0.3863} -0.1927

-0.6105 {0.2255} -2.49046 *

0.1045 {0.4178} -0.2205

H0 0.4855 {0.5193} -0.1736 *** {0.0000} -2.3195 *** {0.0000} -0.1590 {0.9026} 1.4856 {0.4006} -2.1125 **

June 2004 - Dec 2006 (contraction) FedNews H1 H2 H0 0.7655 -0.4120 ** -0.6734 * {0.1724} {0.0308} {0.0875} -4.1458 *** -0.1661 *** 0.0990 ** {0.0000} {0.0000} {0.0263} -3.0860 ** -2.2990 * 2.9791 * {0.0123} {0.0820} {0.0815} 1.5096 -0.8898 1.1608 {0.2122} {0.2290} {0.5380} 2.5768 -0.8534 -0.3259 {0.1218} {0.3827} {0.7817} -2.1124 ** -2.1105 ** 0.4396 *

FedNews H1 -0.8045 ** {0.0336} 0.0056 * {0.0952} 1.0283 {0.6101} 0.6583 {0.7025} -0.1491 {0.9365} 0.5928 *

H2 -0.3850 ** {0.0428} 0.1560 {0.6247} 2.8535 {0.1727} 3.3660 ** {0.0253} 0.3206 {0.8221} 0.7545 *

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

122

Table 5.11 Continued

Panel A Continued
The Feds' target interest rate news spillover effects June 2004 - Dec 2006 (contraction)

Jan 1999 - May 2004 (expansion)

Malaysia

New Zealand

Philippines

Singapore

Taiwan

Thailand

H0 {0.3287} -0.2966 {0.4887} -0.8590 ** {0.0330} 0.2182 {0.2085} -0.5396 * {0.0844} -0.5847 {0.3522} -0.2608 {0.1225}

FedNews H1 {0.0605} -0.4134 {0.1311} -0.2098 ** {0.0127} -0.1061 {0.8465} -1.5130 *** {0.0002} -0.1385 {0.1268} -1.1916 *** {0.0005} H2 {0.2720} -0.0759 {0.8159} 0.0001 {0.9999} 0.2074 {0.1259} 0.4580 {0.2575} -0.1480 *** {0.0004} 0.2611 {0.1208} H0 {0.5106} -0.5634 ** {0.0144} 0.8183 ** {0.0282} 0.0338 {0.8733} 0.2713 {0.3587} 0.2367 {0.2151} 0.6158 * {0.0919} FedNews H1 {0.7063} -1.0990 *** {0.0007} 0.6726 ** {0.0274} 0.1249 {0.7536} 0.0309 {0.9115} 0.1990 {0.3082} -0.6224 {0.3187} H2 {0.4247} 2.0789 ** {0.0172} 0.1105 * {0.0721} 0.0047 {0.9880} 0.7371 * {0.0560} -0.1680 {0.4725} 0.6180 {0.1142} H0 {0.0246} -0.2215 {0.1998} -2.1855 *** {0.0081} -0.7528 {0.5257} 0.6623 {0.2945} -1.3153 {0.3394} -1.0813 {0.2434} H2 {0.0126} -0.2223 {0.1492} -2.2784 * {0.0523} 0.0345 {0.7488} -0.8132 {0.2328} -0.4920 {0.5314} -1.4635 * {0.0883} FedNews H1 {0.0198} -0.0188 {0.9708} -1.4174 * {0.0506} 1.3875 {0.4419} -2.1768 {0.1031} 1.2986 {0.3271} -3.1738 ** {0.0182}

H0 {0.0746} -3.3411 * {0.0588} -0.2303 {0.8676} 3.1688 ** {0.0318} -1.1971 {0.3323} 0.1700 {0.2816} 1.2672 {0.4862}

FedNews H1 {0.0741} -2.9550 {0.1502} -1.6544 {0.2921} 3.5259 *** {0.0081} 0.2518 * {0.0899} 0.1554 {0.2890} -1.5386 {0.4068}

H2 {0.0735} -3.3439 * {0.0556} -0.7801 {0.4985} 0.0809 {0.9459} -0.6015 {0.7609} 0.3059 * {0.0993} 1.1613 {0.5806}

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

123

Table 5.11 Continued

Panel B The ECBs policy cycle impacts

The ECB' target interest rate news spillover effects May 2001 - Nov 2005 (expansion)

Jan 1999 - April 2001 (contraction)

Australia

China

Hongkong

H0 -0.9759 *** {0.0042} -1.0135 * {0.0825} -0.7743 {0.9299}

ECBNews H1 -0.8581 {0.7673} 0.9342 {0.7312} -0.5213 * {0.0947} H2 -0.6628 {0.1515} -0.5113 ** {0.0467} -0.5622 {0.5940} H0 0.0588 *** {0.0001} 0.0507 * {0.0765} 0.6613 {0.3481} ECBNews H1 0.2068 ** {0.0286} 0.0029 *** {0.0000} 0.3051 * {0.0540} H2 0.1580 * {0.0568} 0.5904 ** {0.0492} 0.1554 *** {0.0000}

Indonesia -0.6058 * {0.0703} 0.0673 *** {0.0048} 0.9743 ** {0.0364} 0.0397 * {0.0635}

Japan

-0.9785 ** {0.0417}

-0.1147 {0.2913}

Korea 0.8336 {0.4057} 0.0966 ** {0.0486} 1.2163 *** {0.0033} 0.0966 *** {0.0001}

Malaysia

-0.8161 {0.8729}

-0.3345 {0.5791}

New Zealand

Philippines

Singapore

Taiwan

Thailand

-0.0414 {0.9405} -0.1861 {0.1942} -3.5549 {0.5803} -1.2074 {0.7049}

-0.2013 * {0.0511} -0.9149 * {0.0850} -3.1088 {0.4722} -1.0980 *** {0.0037} 0.8807 {0.3089} -0.4853 {0.2541} -2.5409 ** {0.0199} -0.6685 {0.8457}

5.8334 ** {0.0309} 0.0410 ** {0.0120} 0.1038 *** {0.0000} 0.7182 *** {0.0060}

0.1920 *** {0.0009} 0.1939 *** {0.0000} 0.0478 {0.5518} 0.0585 *** {0.0003}

0.0804 ** {0.0123} 0.0410 ** {0.0349} 0.3476 ** {0.0137} 0.1339 *** {0.0000}

H0 -2.5989 {0.6066} -0.2342 {0.6738} -0.6849 {0.3012} 0.0787 {0.9947} 3.5741 {0.6895} -3.7367 {0.4744} 2.7077 {0.7916} -4.7917 {0.1838} 0.2775 {0.4445} -1.6691 {0.2851} 2.7425 {0.8207} -1.6196 {0.8685}

ECBNews H1 -2.3494 ** {0.0138} -0.6161 {0.2539} -1.7803 {0.2416} 0.4222 {0.6520} 8.9292 {0.2672} -4.4567 ** {0.0377} 1.6601 {0.1917} -1.0313 {0.7995} 0.9739 {0.5484} -1.9656 * {0.0697} 0.5002 {0.1920} -0.5993 {0.8991} H2 -0.0824 {0.8881} -0.1816 {0.6629} -1.6934 ** {0.0247} -1.3848 {0.4140} 2.5659 {0.5697} -3.7688 {0.4532} 1.3458 {0.1017} -0.8169 *** {0.0000} -2.0230 {0.3576} -1.9949 {0.5679} -2.1810 {0.7481} -0.7622 {0.1421}

H0 0.3938 *** {0.0000} 0.9771 {0.8704} 0.1501 {0.1259} 0.2163 {0.1698} 0.7483 {0.8819} -1.7551 {0.1991} 0.0488 *** {0.0000} 0.5458 *** {0.0000} 0.7776 ** {0.0340} 0.3113 *** {0.0014} 0.0446 {0.6948} 0.0756 {0.3729}

ECBNews H1 0.9951 *** {0.0000} 0.7653 {0.2773} 0.0024 *** {0.0000} 0.0526 {0.1638} 0.0656 *** {0.0000} -4.5725 {0.7678} 2.0113 *** {0.0000} 0.0807 *** {0.0000} 0.0178 {0.1861} 0.0829 *** {0.0000} 0.2227 {0.8728} 0.1352 {0.1121}

H2 1.1873 *** {0.0000} 0.7634 {0.7086} 0.2338 *** {0.0000} 3.8899 {0.9396} 0.6571 *** {0.0000} -2.7765 {0.2274} 0.0433 *** {0.0000} 0.4321 * {0.0980} 0.0072 *** {0.0000} 0.0507 *** {0.0000} 0.0120 {0.3384} 0.8774 {0.2398}

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

124

Table 5.12 - The Fed and the ECB asymmetric effects

This table reports the estimation results of the EGARCH models described below for Asia Pacific stock markets daily returns against the overall surprises of the Fed and the ECB. News_R and News_F are unexpected rise (positive in sign) or fall (negative in sign) in the Fed/ECB target rate. P-values are in braces.

y t = [RHS of (5.1a)] + News _ R News _ Rt 1 + News _ F News _ Ft 1

i

ln ht = [RHS of (5.1b)] + News _ R News _ Rt 1 + News _ F News _ Ft 1 + MacroAnn,i MacroAnni ,t 1

(5.8a) (5.8b)

Panel A The Feds asymmetric effects

FedNews_R

FedNews_F
***

**

FedNews_R

FedNews_F

H0 0.6510 {0.5956} -0.2378 {0.2341} 0.0286 {0.9700} 0.0496 {0.8017}

Australia H1 0.9663 {0.1284} -0.1867 {0.3032} -0.2964 {0.7287} -0.0297 {0.9061} H0 -0.6315 {0.6060} -0.2011 {0.4811} 0.6861 {0.4668} 0.2123 {0.5786} H2 0.4008 {0.1452} -0.7508 * {0.0959} -0.5201 {0.6461} -0.3142 * {0.0945}

H2 -0.0797 {0.7008} 0.0307 {0.6676} -0.1959 {0.0007} -0.3377 {0.5482}

H0 -0.3225 {0.8146} -0.8074 {0.0294} 0.8800 {0.6057} 4.2326 {0.2607}

Indonesia H1 1.4486 {0.3107} 3.5864 {0.1180} 0.3019 {0.8874} 4.2174 {0.2814}

**

H2 -0.9158 {0.2193} -0.6457 {0.6050} 3.4471 {0.3648} 2.6054 {0.8352}

Mean Eq
***

Var Eq

0.5116 {0.4744} 0.0007 {0.9784} 0.1237 {0.7250} 0.2222 {0.6374}

3.0042 {0.0830} 0.1014 {0.7502}

0.2519 {0.6158} 8.7290 {0.0031}

China Hongkong H1 H2 H0 H1 -1.5109 1.6870 *** 0.5868 0.7848 * {0.3401} {0.0034} {0.1580} {0.0568} -0.2085 -0.2178 *** -0.1069 -0.4673 *** {0.4910} {0.0000} {0.6744} {0.0000} 0.6347 0.1078 *** -0.1313 *** -0.2702 * {0.4790} {0.0000} {0.0049} {0.0892} 0.4169 -0.7721 *** -0.2316 ** -0.1895 {0.3517} {0.0000} {0.0297} {0.3693} Tests of equality between FedNews_R and FedNews_F 0.6590 0.9279 *** 2.1506 8.2421 *** {0.4169} {0.0009} {0.1425} {0.0041} 0.0519 2.3352 *** 5.0941 ** 2.0845 {0.8197} {0.0000} {0.0240} {0.1488} 4.5201 ** {0.0335} 0.0341 {0.8536}

4.4565 {0.0348} 0.6919 {0.4055}

0.5179 {0.4717} 0.8241 {0.3640}

0.0280 {0.8671} 0.0033 {0.9543}

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

125

Table 5.12 Continued

FedNews_R
***

***

FedNews_F

FedNews_R

FedNews_F

H0 0.5704 {0.7425} -0.4676 {0.2292} -0.5279 {0.5866} -0.2099 {0.2636}

***

Japan H1 2.2466 {0.1981} -0.6291 {0.3637} -0.3204 {0.8227} 0.9812 {0.3063} H0 -0.1122 {0.0228} 0.5233 {0.6752} 2.2444 {0.0077} -0.0508 {0.8367} H2 -0.4666 {0.1126} 0.7296 {0.2739} -1.9300 {0.2658} 2.4532 * {0.0650} H0 -0.6018 {0.0015} -1.1759 {0.0003} 0.7586 {0.5021} 0.4650 {0.1186}

H2 -0.9447 {0.2128} 0.2190 {0.2993} -0.0334 {0.9612} 0.1027 {0.6381}

Malaysia H1 -0.0848 *** {0.0000} -0.4563 ** {0.0401} -0.3607 {0.7482} -1.0465 *** {0.0000}

New Zealand H1 -0.3080 *** {0.0060} -2.3190 *** {0.0017} 0.7374 {0.4488} 0.2598 {0.4103}

H2 -0.3754 *** {0.0000} 0.0013 {0.9966} 0.7245 *** {0.0000} 0.8962 *** {0.0038}

Mean Eq

Var Eq

0.3560 {0.5508} 0.1033 {0.7479}

2.4197 {0.1198} 0.5947 {0.4406}

2.1339 {0.1441} 0.0342 {0.8533}

3.6666 {0.0555} 6.2395 {0.0125}

Korea H1 H2 H0 ** -0.7278 *** -4.1499 ** -0.3991 *** {0.0054} {0.0113} {0.0000} -2.7639 ** 0.5819 -0.0471 {0.0225} {0.6169} {0.8901} *** 1.9565 ** 2.0542 ** -0.7265 {0.0432} {0.0356} {0.2286} -0.0976 -0.0282 -0.6949 *** {0.7213} {0.9169} {0.0000} Tests of equality between FedNews_R and FedNews_F * 12.8642 *** 5.3642 ** 1.0678 {0.0003} {0.0206} {0.3014} ** 3.8360 * 4.4316 ** 0.0023 {0.0502} {0.0353} {0.9619} 2.8215 * {0.0930} 0.3500 {0.5541} 2.8759 * {0.0899} 3.4809 * {0.0621} 17.0931 {0.0000} 1.1004 {0.2942} 17.3790 *** {0.0000} 1.0352 {0.3089}

1.4224 {0.2330} 3.8669 *** {0.0000}

FedNews_R

***

**

FedNews_F
***

FedNews_R

***

FedNews_F

H0 -0.3243 {0.0048} 0.2329 {0.1585} 0.4835 {0.0000} 0.1234 {0.8551}

**

Philippines H1 0.2076 {0.4042} -0.0959 {0.8408} 0.0891 *** {0.0012} 0.1295 {0.7478} H0 0.7677 {0.1573} -0.4056 {0.3613} 0.5050 {0.0009} 0.0789 {0.7486}

H2 -0.4765 {0.0466} 0.2370 {0.1063} 0.5530 {0.0000} 0.1304 {0.5472}

Taiwan H1 0.6374 {0.1646} -0.3530 {0.1520} 0.2350 *** {0.0039} 0.0847 {0.6085}

H2 -0.9893 {0.2348} -0.0919 *** {0.0063} 0.3204 ** {0.0341} -0.1590 {0.6556}

H0 -0.2163 {0.1707} 0.2514 {0.1413} -3.3143 {0.1725} 0.2233 {0.8154}

*

Thailand H1 -0.7599 *** {0.0003} -0.6327 {0.1806} -0.5261 {0.7379} 0.2959 {0.4560}

H2 -0.6352 {0.5976} 0.2091 {0.2387} 1.7228 {0.7129} -0.4980 {0.3831}

Mean Eq
***

***

Var Eq

0.9245 {0.0028} 1.4311 {0.0000} 2.9775 {0.0844} 11.0499 {0.0009}

***

0.7865 {0.3752} 9.4382 {0.0021}

***

4.3615 {0.0368} 34.1624 {0.0000}

Singapore H1 H2 H0 -0.4810 *** -0.7236 *** 0.1446 {0.0000} {0.0034} {0.1176} -0.4233 *** -0.7676 ** -0.7944 {0.0001} {0.0273} {0.2743} *** 0.1245 ** 0.6351 0.0797 ** {0.0319} {0.6406} {0.0180} -0.2187 0.4872 ** 0.1422 {0.2640} {0.0451} {0.4288} Tests of equality between FedNews_R and FedNews_F * 17.6692 *** 15.8180 *** 3.6215 * {0.0000} {0.0001} {0.0570} *** 3.6634 * 0.7351 5.8323 ** {0.0556} {0.3912} {0.0157} 4.4717 ** {0.0345} 8.1918 *** {0.0042}

5.2042 ** {0.0225} 4.1915 ** {0.0406}

3.5905 {0.0581} 1.8736 {0.1711}

3.2939 *** {0.0003} 0.2537 {0.6145}

0.4748 {0.4908} 0.1996 {0.6551}

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

126

Table 5.12 Continued

Panel B The ECBs asymmetric effects

ECBNews_R
*** *

ECBNews_F

ECBNews_R

ECBNews_F

H0 -0.3496 {0.9296} -2.8694 {0.1843} 2.8945 {0.3956} 2.3549 {0.1814}

Australia H1 2.6930 {0.4628} -1.8612 {0.4192} -0.4654 {0.8780} -0.3588 {0.8576} H0 -5.9063 {0.4529} 1.6206 {0.6463} -2.0083 {0.5459} -5.0891 {0.0039} H2 -8.2130 *** {0.0100} -0.5639 {0.8331} -0.5491 {0.8510} -2.6994 {0.1635} H0 1.7478 {0.8163} -0.7059 {0.0745} -9.4428 {0.8057} -10.7425 {0.7244}

H2 -1.5226 {0.1901} -0.3530 {0.0000} 7.0950 {0.3590} -2.8561 {0.3988}

Indonesia H1 -0.7191 ** {0.0434} -0.6245 *** {0.0003} -22.5927 {0.6569} -22.7492 {0.3307}

H2 -4.6991 {0.5087} -0.0208 *** {0.0000} 9.0588 {0.1159} 6.3590 *** {0.0004}

Mean Eq

Var Eq

0.3527 {0.5526} 0.0244 {0.8758}

1.1125 {0.2915} 0.0013 {0.9717}

1.0128 {0.3142} 1.8239 {0.1769}

0.7676 {0.3810} 0.7005 {0.4026}

China Hongkong H1 H2 H0 H1 -5.3419 * 0.2419 6.8657 14.7061 {0.0796} {0.3940} {0.2734} {0.1410} -0.8051 1.5717 -6.0753 ** -5.2567 {0.7311} {0.1324} {0.0313} {0.2668} -5.0647 -2.2987 *** 0.7162 * 0.5770 {0.1264} {0.0000} {0.0871} {0.7825} *** -5.3631 -4.5352 -1.0471 -0.6350 {0.1024} {0.3209} {0.3399} {0.7423} Tests of equality between ECBNews_R and ECBNews_F 3.0531 * 0.3701 1.4496 3.5264 * {0.5430} {0.2286} {0.0604} {0.0806} 0.0011 7.6587 *** 0.9899 0.2305 {0.9738} {0.0056} {0.3198} {0.6311} 1.6383 {0.2006} 0.1883 {0.6643} 0.3130 {0.5758} 0.0038 {0.9508} 5.7077 ** {0.0169} 1.8822 {0.9965}

0.0090 {0.9243} 0.1914 {0.6617}

ECBNews_R

ECBNews_F
***

***

ECBNews_R
**

ECBNews_F

H0 0.5625 {0.3347} -0.7482 {0.8852} 1.0266 {0.7324} 4.0787 {0.0028}

***

Japan H1 0.6306 {0.3400} -0.3408 {0.9319} 2.9018 {0.3297} 4.5236 {0.0262} H0 -1.5093 {0.8626} -3.8115 {0.5511} -6.8823 {0.1685} -2.2993 {0.2156}

**

H2 0.0864 {0.8925} -0.5149 {0.7952} 0.3340 {0.0094} 0.4133 {0.0376}

H2 0.4146 {0.6776} -0.1940 {0.4715} 4.5319 ** {0.0236} -1.2491 {0.1015}

H0 -5.0083 {0.4546} -3.4452 {0.0000} 3.8949 {0.8029} 3.4887 {0.2423}

New Zealand H1 -5.3762 {0.4739} -1.5214 {0.6566} 4.0144 {0.4781} 2.5127 {0.2492}

H2 -0.1959 {0.6543} -5.8697 *** {0.0000} 1.2503 {0.4835} 4.5425 *** {0.0000}

Mean Eq
***

Var Eq

0.6230 {0.4299} 0.9956 {0.3184}

0.7323 {0.3921} 0.2952 {0.5869}

0.0841 {0.7719} 0.9276 {0.0006}

0.0448 {0.8323} 0.8701 {0.3509}

Korea Malaysia H1 H2 H0 H1 3.7059 -1.6476 -0.6548 ** -0.7642 * {0.8406} {0.9204} {0.0398} {0.0865} -3.0696 *** -3.8277 -0.9995 ** -0.0350 * {0.0052} {0.4743} {0.0485} {0.0798} -8.6803 -6.9366 4.8670 ** 2.5945 {0.2211} {0.2892} {0.0338} {0.4587} -4.0859 * -2.3222 -3.9019 ** -1.6765 ** {0.0735} {0.3433} {0.0120} {0.0204} Tests of equality between ECBNews_R and ECBNews_F 0.1180 0.0159 7.4497 *** 5.1041 ** {0.7313} {0.8996} {0.0063} {0.0239} 0.4509 0.4864 12.5040 *** 4.1497 ** {0.5019} {0.4855} {0.0004} {0.0416}

0.3887 {0.5330} 0.3118 {0.5766}

1.5528 {0.2127} 0.0005 {0.9817}

0.6483 {0.4207} 0.0676 {0.7948}

4.3580 ** {0.0131} 3.1967 {0.7119}

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

127

Table 5.12 Continued

**

ECBNews_R
* *

ECBNews_F
***

***

ECBNews_R
*** ***

ECBNews_F

H0 0.9738 {0.3698} 0.6069 {0.0000} -7.0663 {0.2974} 3.6081 {0.0000}

**

***

Philippines H1 6.9373 {0.1537} 1.8330 {0.8514} -7.9463 {0.2224} 9.5085 ** {0.0417} H0 -0.3623 {0.0052} 0.6955 {0.1611} 0.3545 {0.0223} 0.1576 {0.3868} H2 -3.8283 {0.3246} -1.2582 {0.6351} -1.4833 {0.6436} -1.5148 {0.4572} H0 -1.0878 {0.0295} -2.3844 {0.0706} 3.0448 {0.0000} 0.2778 {0.0069}

H2 0.7908 {0.7614} 0.6144 {0.0862} -5.5830 {0.8025} 3.4404 {0.0000}

Thailand H1 0.5798 {0.2763} 0.9349 {0.8159} 1.1248 {0.1995} 0.2370 {0.1949}

H2 -0.1133 ** {0.0214} 2.3788 {0.1167} 2.9068 *** {0.0006} 0.3194 {0.2592}

Mean Eq

Var Eq

0.1107 {0.7393} 7.7744 {0.0053}

***

1.6394 {0.2004} 4.7414 {0.0294}

**

0.0046 {0.9457} 1.5379 {0.2149}

1.3840 {0.2394} 2.4603 {0.1168}

Singapore Taiwan H1 H2 H0 H1 *** -0.7460 * 0.3873 -2.6928 2.1959 {0.0568} {0.7721} {0.6686} {0.7763} 0.2356 0.6271 -3.7803 -2.7490 {0.6452} {0.3110} {0.4313} {0.5390} ** 0.7044 0.5706 *** 4.1872 0.7997 {0.2690} {0.0015} {0.2581} {0.8428} 0.0457 * 0.0989 ** 0.2168 -0.3969 {0.0570} {0.0200} {0.8940} {0.8544} Tests of equality between ECBNews_R and ECBNews_F 1.5169 0.0253 0.0188 0.2699 {0.2181} {0.8737} {0.8909} {0.6034} 0.1072 2.2710 1.2385 0.0672 {0.7434} {0.1318} {0.2658} {0.7955} 0.2953 {0.5869} 0.0001 {0.9932} 6.4033 {0.0114} 8.7091 {0.0000}
***

0.7095 {0.3996} 3.8279 * {0.0504}

6.6852 *** {0.0097} 5.8242 *** {0.0001}

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

128

5.5. Conclusion This chapter documents and discusses the existence and the nature of the spillover effects of the policy interest rate news from the Fed and the ECB on the first two moments of the market returns of twelve Asia Pacific stock markets for the period January 1999 to December 2006. Considering that these are the two most important economic blocs in the world, and their monetary policy shocks have been reported to elicit significant stock market movements not only in their respective markets but also in others, the comprehensive investigation of the spillover effects of these interest rate news in the stock markets of the Asia Pacific region would enrich the current literature. In general, it has been revealed that the spillover news effects of both the Feds and ECBs news on the Asia Pacific market returns are generally consistent with the literature where a majority of stock markets show negative returns in response to unexpected rate hikes, and there is some evidence that the ECB news is absorbed slowly. This might be to some extent due to a supplementary role of the ECBs news for the Feds news in the Asia Pacific stock markets. More importantly, significant volatility responses to the interest rate news from the two central banks has been documented and discussed in details. This study is the first to document such comprehensive empirical evidence. Unexpected hikes in the Feds and the ECBs target rates increased the return volatility in most of the Asia Pacific stock markets. Even though the ECBs interest rate decisions may have been influenced by those of the Fed, the ECB news was also an important informational provider in the Asia Pacific region. In addition, in general, delayed
129

volatility responses have been found. In most of the markets, the volatility response was significant only in the intradaily trading hours. In other cases, both interest rate news elicited a significant volatility response in the overnight horizon that persisted into the intradaily trading hours. These findings have important implications for the Asia Pacifics policy makers and market participants alike in anticipating and managing potential spillover effects from the Feds and the ECBs so as to have forward looking policy actions and hedging strategies. For instance, the notion that the ECB follows the Fed in setting its policy is so strong amongst market participants, there might be argument that policy makers and market participants need not to consider the ECBs target rate actions but keep a close watch at the Fed only. We, however, provided evidence that the ECB news was also another important informational leader that Asia Pacific region should follow closely as the ECBs impacts on the Asia Pacific markets are consistent with those imposed by the Feds decisions. Furthermore, understanding that unexpected hikes in the Feds and the ECBs target rates would result in a negative return in most of the Asia Pacific stock indices, index traders would also design their trading and hedging strategies accordingly to shield the risk and make profits if their expectations are correct.

130

CHAPTER 6 - TARGET INTEREST RATE NEWS EFFECTS ON THE ASIA PACIFIC CURRENCY MARKETS44

44

A shorter version of this chapter is published in International Research Journal of Finance and

Economics, 20, 27-45, 2008.

131

6.1. Introduction Aside from the interest rate and the stock markets, currency markets might also be influenced by monetary policy. The empirical literature in this area has developed broadly along two branches: a VAR approach and an event-study approach. Using of VAR models, Lewis (1995) and Eichenbaum and Evans (1995) find that the initial responses of the exchange rate to various measures of monetary policy news are small or insignificant, and that such news has their largest impact on exchange rates several months after the initial shock.45 However, Rudebusch (1998) and Brunner (2000) raise several questions related to the causal structure and reliability of VAR-based models for measuring policy news. Faust and Rogers (2003) and Faust et al. (2003) re-examine the issue of monetary policy effects on exchange rates using an inference procedure that relaxes the strong identifying assumptions such as foreign interest rates do not respond to the Fed policy moves until a month after they are made, e.g. Eichenbaum and Evans (1995) and Kim and Roubini (2000)s finding of delayed overshooting effects of the Feds news on the British Pound and German Mark/Euro, and such policy news can explain only a small portion of currency variability. On the other hand, event-study researches find stronger evidence of an immediate impact of monetary policy news on exchange rates. Andersen et al. (2003) and Ehrmann and Fratzscher (2005b) show that the USs and the euro areas (and Germans news before 1999) macroeconomic news have a significant impact on

45

A possible explanation why these authors do not find evidence for an immediate effect is that they

do not distinguish between news announcements and surprise component,

132

the US dollar-euro exchange rate. Fatum and Scholnick (2008) find that the unexpected component of a tightening (loosening) of US monetary policy is associated with a same day appreciation (depreciation) of the US dollar against the British Pound, German Mark, and Japanese Yen. Furthermore, Lobo et al. (2006) report that changes in the Feds interest rate target are positively related to changes in the value of the dollar, and also provide evidence for the asymmetric impact of the Feds news where unexpected rate rises have a larger effect than unexpected rate cuts for the Pound, Mark, and Canadian dollar, whereas the opposite is true for the Japanese yen. Generally, the large body of evidence suggests that the Feds monetary policy actions provide useful information to currency markets. This chapter aims to provide comprehensive evidence on the existence and the nature of the spillover effects of both the Feds and the ECBs policy interest rate news on the first two moments of exchange rates of the Asia Pacifics currencies against the US dollar (USD) and the euro (EUR) for the period from January 1999 (since the introduction of the ECB) to December 2006. The key findings of this chapter are summarized as follows. First, the spillover effects of the Feds and the ECBs news on the Asia Pacific currencies are generally consistent with the literature and the arbitrage arguments, where a majority of currencies depreciate against the USD and the EUR in response to an unexpected hike in the Feds and ECBs target rates. Second, unexpected hikes in the Feds and the ECBs target rates reduced volatility in most of examined currency pairs. Third, both sets of news tend to be absorbed slowly in the Asia Pacific currency markets. Fourth, even though the ECBs interest rate decisions may have been influenced by those of the Fed, the ECBs news is also an important informational provider in the Asia Pacific region. The Feds news seems to send a leading signal upon the upcoming decision of the
133

ECB, while the ECBs news tends to confirm the Feds decision. This relationship between the two policy rate news helps to reduce volatility in the Asia Pacific currency markets. The rest of the chapter proceeds as follows. Section 6.2 discusses data requirements, and the empirical modeling issues are discussed in Section 6.3. Section 6.4 reports and analyzes the estimation results, and Section 6.5 concludes the chapter.

6.2. Data The news series used in this investigation is presented in Table 3.1 in page 35. Data on the Asia Pacific currencies were collected from Bloomberg. Daily open and close spot foreign exchange rates for seven currencies have been collected, namely Australian dollar (AUD), New Zealand dollar (NZD), Japanese yen (JPY), Korean won (KRW), the Pilipino peso (PHP), Indonesian rupiahs (IDR), and Taiwanese dollar (TWD), against the USD and the EUR for the period from January 1999 to December 2006.46 Daily open and close prices allow for an investigation of the adjustment speed of the Asia Pacific currency markets to the news from the Fed and the ECB. In order to do so, the overall daily horizon (H0) is decomposed into two sub-horizons,

46

The sample exchange rates were indirectly quoted as 1 unit of Asia Pacific currencies equals x units

of the USD or the EUR. Therefore, if the USD or the EUR is strengthening (i.e., appreciating) then the exchange rate number decreases. Due to data limitation, daily open and close data was only available for the full sample period for AUD, NZD and JPY exchange rates against the USD and the EUR only. The others have different starting dates as noticed in Table 6.1.

134

overnight (H1) and intradaily (H2). The daily horizon (H0) is from market close of day t-1 to day t enveloping the Feds news on calendar day t, which is the focus of the previous studies. The overnight horizon (H1) is from the close of day t-1 to the open on day t and this captures the first reaction to the Feds and the ECBs news. The intraday horizon (H2) is from opening of day t to closing of day t and this captures any delayed reactions during the next trading day in Asia Pacific local markets.47 Table 6.1 reports the summary statistics of the exchange rates changes over the three horizons. There are significant differences across the three time horizons in various aspects. The USD exchange rates show negative returns over H1 and positive returns over H2 in all cases except for the PHP/USD pair. For H0, there is a mix of positive and negative returns (4 and 3 cases, respectively). For exchange rates against the EUR, in most cases except for the AUD/EUR and NZD/EUR pairs, the series have negative means during the daily (H0) and intraday (H2) horizons, while more positive means are observed overnight (H1). Noticeably, in three cases the series have all negative means over all three horizons (IDR/EUR, JPY/EUR, and TWD/EUR). In all cases, the return volatility measured by the standard errors is significantly lower during the overnight than during the other horizons. AUD/USD and NZD/USD pairs show negative skewness over H0 and H2 and positive skewness over H1, while the reverse is true for JPY/USD, PHP/USD and TWD/USD pairs. The return series of exchange rates against the EUR are negatively skewed over the

47

This is equivalent to same-day effect identified by Fatum and Scholnick (2008).

135

daily and intraday horizons for four pairs of AUD/EUR, IDR/EUR, KRW/EUR and NZD/EUR, and positively skewed overnight, while the reverse is true for PHP/EUR. JPY/EUR shows positive skewness in all three time horizons while the reverse is observed in the TWDEUR pair. In general, the series exhibit leptokurtosis. 6.3. Empirical methodology 6.3.1. Baseline EGARCH(1,1) model This investigation starts with the baseline univariate EGARCH(1,1) model and then progress to specific modeling of the interest rate news effects in various forms of the Feds and the ECBs news on the Asia Pacific currencies. The baseline EGARCH(1,1) model employed in this study is described by the conditional mean and the conditional variance equations (6.1a) and (6.1b) shown below.
y t = c + Lag y t 1 + Hol Mont + t

(6.1a) (6.1b)

ln ht = c + h ln ht 1 + 1

t 1
ht 1

+ 2

t 1
ht 1

+ Hol Mont

The conditional mean equation for the returns ( yt ) is expressed as a function of past returns as well as Monday effect (Mon) in relevant markets. The conditional variance equation for the returns in the financial market series ( ht ) is expressed as a function of the one period lag of the variance, Monday effect (Mon), and the residuals. The Monday effect variable takes the value of one for Monday and zero otherwise to control for the weekend effect where more intense information flows might show up after a period of market closure.

136

Table 6.1 - Descriptive statistics of the Asia Pacific currency markets returns

This table reports descriptive statistics for the Asia Pacific exchange rates changes for the period from January 1999 to December 2006. H0 is daily horizon over the close on calendar day t-1 to the close on calendar day t; H1 is overnight horizon over the close of day t-1 to the open on day t; and H2 is intraday horizon measured over the open on day t to the close on day t. P-values are in braces.

Panel A. Asia Pacific currencies/USD

H2 2085 0.0174 0.6689 -0.2908 1.1536 145.07
*** *** *** ***

H0 1389 -0.0004 0.6189 -0.4591 10.5740 6519.80 1389 0.0345 0.6423 -0.0586 11.3461 7456.68

H0

H2

2085 0.0123 0.6689 -0.2541 1.1072 128.94

AUDUSD H1 2085 -0.0053 0.1001 1.2086 29.2856 75015.51 H0 2085 -0.0021 0.6234 0.0790 2.0907 381.91 H2 2085 0.0033 0.6129 0.0283 2.1054 385.55

IDRUSD H1 1389 -0.0358 0.3078 -0.7082 16.2073 15318.61

JPYUSD H1 2085 -0.0051 0.1380 -0.9661 63.2502 347875.11

H0 1387 0.0230 0.4352 -0.1637 2.8074 461.68

KRWUSD H1 1387 -0.0155 0.2916 -0.4850 6.6423 2604.19

H2 1387 0.0383 0.4684 0.0104 2.7025 422.41

***

No. of observations Mean Std. deviation Skewness Kurtosis Jarque-Bera Ljung-Box Q test Return P-Value Volatility P-Value
***

24.2256 {0.2327} 171.9142 {0.0000}

*** *** *** ***

***

52.3656 {0.0001} 133.5441 {0.0000}

22.0267 {0.3391} 161.8804 {0.0000}

54.7285 {0.0000} 348.3105 {0.0000}

565.6338 {0.0000} 243.6084 {0.0000}

89.2077 {0.0000} 343.5190 {0.0000}

26.7803 {0.1416} 108.2925 {0.0000}

76.7875 {0.0000} 83.9396 {0.0000}

***

15.0313 {0.7746} 82.9997 {0.0000}

***

18.1890 {0.5750} 32.4102 ** {0.0391}

583.3748 *** {0.0000} 787.8080 *** {0.0000}

123.5925 {0.0000} 191.1708 {0.0000}

***

H0 2085 0.0184 0.7285 -0.2874 0.9798 112.16

*** ***

H2 1824 -0.0112 0.4574 3.0975 48.7117 183251.43 1824 -0.0174 0.5121 6.5923 122.0836 1146573.95

H0

H2

2085 0.0137 0.7307 -0.2613 0.9667 104.90

NZDUSD H1 2085 -0.0046 0.1170 0.1256 16.0833 22477.62 H0 1894 -0.0013 0.2624 0.1598 10.4079 8556.72

PHPUSD H1 1824 0.0065 0.3886 -10.9757 300.0882 6880643.28

TWDUSD H1 1894 -0.0088 0.2075 -1.9780 34.7862 96730.43

**

H2 1894 0.0075 0.2739 0.2985 7.1881 4107.86

***

No. of observations Mean Std. deviation Skewness Kurtosis Jarque-Bera Ljung-Box Q test Return P-Value Volatility P-Value
***

18.8665 {0.5305} 152.7418 {0.0000}

*** ***

***

15.4074 {0.7526} 365.7286 {0.0000}

19.6633 {0.4792} 144.3772 {0.0000}

38.5794 {0.0075} 207.1359 {0.0000}

27.1940 {0.1299} 0.3103 {1.0000}

50.9597 {0.0002} 8.3327 {0.9894}

35.4462 {0.0179} 167.2696 {0.0000}

***

1134.0070 {0.0000} 23.4800 {0.2658}

547.5142 {0.0000} 77.2777 {0.0000}

***

Note: Data starting dates: IDRUSD (3/9/2001), KRWUSD (5/9/2001), PHPUSD (3/1/2000), and TWDUSD (27/9/1999). All the others are from Jan 1999. *, **, *** denotes significance at 10%, 5%, and 1%, respectively.

137

Table 6.1 - Continued

Panel B. Asia Pacific currencies/EUR

H2 2085 -0.0018 0.6448 -0.3781 2.2789 501.08
*** ** *** ** ***

H0 1158 -0.0247 0.7415 -0.2498 1.9437 194.32 1158 -0.0137 0.7598 -0.2197 1.8160 168.58

H0

H2

2085 0.0065 0.6562 -0.2843 2.3828 521.33

AUDEUR H1 2085 0.0083 0.1663 3.0794 51.4306 233089.02 H0 2085 -0.0083 0.7103 0.0241 2.8565 709.04 H2 2085 -0.0063 0.7017 0.0265 2.9400 751.52 H0 1044 -0.0039 0.6899 -0.0323 0.4781 11.09

IDREUR H1 1158 -0.0114 0.1967 3.0067 42.7844 90066.75

JPYEUR H1 2085 -0.0021 0.1676 2.3067 77.0702 517869.81

KRWEUR H1 1044 0.0013 0.1257 5.8225 112.8419 613417.19

H2 1044 -0.0050 0.6944 -0.0972 0.8352 35.08

***

No. of observations Mean Std. deviation Skewness Kurtosis Jarque-Bera Ljung-Box Q test Return P-Value Volatility P-Value
***

26.2149 {0.1588} 148.7215 {0.0000}

*** *** *** *** *** ***

***

89.7700 {0.0000} 219.7319 {0.0000}

26.3863 {0.1534} 122.4143 {0.0000}

37.3231 {0.0107} 83.2778 {0.0000}

132.2189 {0.0000} 176.2092 {0.0000}

34.8255 {0.0211} 83.9080 {0.0000}

20.1226 {0.4503} 413.4854 {0.0000}

74.7475 {0.0000} 39.7125 {0.0000}

19.3987 {0.4961} 390.5683 {0.0000}

***

75.0859 *** {0.0000} 67.6846 *** {0.0000}

28.7313 * {0.0932} 3.0316 {1.0000}

75.2980 {0.0000} 45.4879 {0.0009}

***

H0 2085 -0.0056 0.7062 -0.3008 1.7585 300.24

** *** *** ** **

H2 1248 -0.0288 0.7147 0.0214 4.1668 902.92 1248 -0.0440 0.6593 0.1071 3.2035 536.46

H0

H2

2085 0.0084 0.7180 -0.1966 1.9779 353.32

NZDEUR H1 2085 0.0135 0.1678 4.0705 45.3494 184422.15 H0 1894 -0.0136 0.6653 -0.0344 0.8974 63.93

PHPEUR H1 1248 0.0156 0.4488 -0.6624 15.8595 13170.56

TWDEUR H1 1894 -0.0107 0.2086 -4.9031 206.6283 3376956.73

***

H2 1894 -0.0030 0.6648 -0.0093 1.3437 142.58

**

No. of observations Mean Std. deviation Skewness Kurtosis Jarque-Bera Ljung-Box Q test Return P-Value Volatility P-Value 23.6223 {0.2593} 114.5481 {0.0000}
*** ***

29.7838 {0.0734} 139.7547 {0.0000}

***

33.9982 {0.0261} 27.0443 {0.1340}

62.8212 {0.0000} 182.4217 {0.0000}

44.2079 {0.0014} 25.4370 {0.1852}

32.3721 {0.0395} 17.3729 {0.6286}

47.5402 {0.0005} 100.0221 {0.0000}

***

168.8278 {0.0000} 5.1628 {0.9996}

31.8769 {0.0446} 46.1953 {0.0008}

***

Note: Data starting dates: IDREUR (23/7/2002), KRWEUR (12/8/2002), PHPEUR (19/3/2002), and TWDEUR (10/02/2000). All the others are from Jan 1999. *, **, *** denotes significance at 10%, 5%, and 1%, respectively.

138

6.3.2. Target interest rate news spillover effects To investigate the spillover impacts of the Feds and the ECBs news on the Asia Pacific currencies returns and volatilities, the baseline model is extended with an overall news variable for the Fed and the ECB, Newst (FedNewst for the Fed and
ECBNewst for the ECB separately). The Newst variable is lagged by one period to

account for the time difference between the US and European and the Asia Pacific trading hours. Furthermore, a dummy variable (MacroAnni,t) for each of the major macroeconomic announcements described in Chapter 5 is added in the conditional variance equation in order to isolate the influence of the target interest rate news on the days of multiple information releases. The extended model is then shown in equations (6.2a) and (6.2b) below. (6.2a) (6.2b)

y t = [RHS of (6.1a)] + News News t 1

ln ht = [RHS of (6.1b)] + News Newst 1 + MacroAnn,i MacroAnni ,t 1

i

In general, the sign of the news coefficient in the mean equation, News, can be either negative or positive. The USD and the EUR would appreciate against other currencies in response to unexpected hike in interest rates in these two economies as reported in the literature, thus a negative sign is observed. However, as Vickers (2000) finds that the reactions of exchange rates to policy announcements could transmit important insights about the markets perception regarding the central banks credibility in fighting inflation. For instance, an unexpected hike in Feds and

139

the ECBs target rates could be viewed as a signal of future economic contraction and lowering the likelihood of future inflation. 48 As a result, such unexpected tightening may lead to a depreciation of the USD and the EUR. In addition, higher interest rates would dampen stock market activities leading to portfolio outflows and this would lead to depreciation of the domestic currency. Therefore, one may observe positive mean coefficients instead. Furthermore, depending on the likely policy response of foreign central banks to the Feds and the ECBs policy decisions, one may even not be able to observe any change in the exchange rates. That is, if market participants believe that the counterpart central banks are more likely to respond to the Feds and the ECBs policy choices, e.g. the counterpart central banks may also unexpectedly increase their own policy interest rates in response to the Feds and the ECBs unexpected actions, the Feds and the ECBs news effect on exchange rates against these correspondent currencies might be muted. The news effect on the conditional volatilities would depend on whether the news adds to or resolves uncertainties in the markets. If an unexpected change in the rate leads to further speculation in the market regarding the future direction of the target rate, this increased heterogeneity would have been shown in a positive news coefficient in the variance equation, i.e. a rise in the volatility. On the other hand, a market calming effect could be observed if an unexpected rate change announcement resolves uncertainty, and a negative news coefficient is shown. In addition, the response would depend on the informational role that the Feds and the ECBs target rate news play in the Asia Pacific markets.

48

This is particularly true in the case of the ECB under inflation targeting framework.

140

6.4. Empirical results 6.4.1. Baseline models results The baseline models results are summarized in Table 6.2. In general, the Monday dummy tends to reduce changes in the exchange rates against the USD, while it significantly increases changes in the exchange rates against the EUR mostly overnight. In the conditional variance equations, the lagged variance term (h) is close to one in most cases suggesting volatility persistence as found in the literature. There is evidence for an asymmetric influence of the past innovations as shown by a negative 1 in most of the cases for the daily and overnight horizons respectively. The volume effect is positive ( 2 >0) in most cases, except for NZD/USD and KRW/EUR overnight rates, indicating that regardless of sign, the higher is the previous periods unexpected movements, the higher its impact on the conditional variance in the next period. On Mondays, in general, the conditional volatility is lower during the intraday and the daily horizons while the reverse is observed overnight for exchange rates against the USD and the EUR. The serial correlations in the first and the second moments of the standardized residuals are removed in all cases by adding more lags of the dependent variable in the mean equation and by varying the lag structures of the EGARCH model in the variance equations. However, the same model specification has been adopted as the news coefficients from these estimations are qualitatively the same as what reported here.

141

Table 6.2 - EGARCH(1,1) estimations of daily returns

This table reports the estimation results of the EGARCH (1,1) model described in (6.1a) and (6.1b) for daily changes series of Asia Pacific exchange rates against the USD and the EUR. H0 is daily horizon over the close on calendar day t-1 to the close on calendar day t; H1 is overnight horizon over the close of day t-1 to the open on day t; and H2 is intraday horizon measured over the open on day t to the close on day t. P-values are in braces.

y t = c + Lag y t 1 + Hol Mont + t

ln ht = c + h ln ht 1 + 1 ht 1
JPYUSD H1 H2
***

t 1
+ 2 ht 1 + Hol Mont

t 1

(6.1a) (6.1b)

Panel A. Asia-Pacific currencies/USD AUDUSD H0 H1 H0

***

H2

H0
**

KRWUSD H1
**

H2
***

c -0.0072 {0.4310} 0.0439 {0.0720}

* ** *** *** *** ***

***

Lag, 1

0.0149 {0.2841} -0.0126 {0.5570}

***

-0.0039 {0.0000} 0.0256 {0.0000}

***

0.0229 {0.0043} -0.0111 {0.4613}

-0.0034 {0.0034} 0.0493 {0.0000}

0.0052 {0.7109} -0.0300 {0.1639}

***

0.0223 {0.0431} -0.0434 {0.1121}

0.0100 {0.0167} 0.2753 {0.0000}

***

***

0.0206 {0.0000} 0.1750 {0.0000}

***

***

c
*** *** ***

***

***

***

h
***

***

***

***

***

***

***

***

1
*** ***

***

**

-0.0742 {0.0000} 0.9859 {0.0000} -0.0114 {0.3602} 0.0798 {0.0000}

***

***

-0.0113 {0.0000} 0.99996 {0.0000} 0.0064 {0.1135} 0.0208 {0.0000}

***

-0.0729 {0.0000} 0.9855 {0.0000} -0.0104 {0.3034} 0.0779 {0.0000}

-0.5234 {0.0000} 0.8129 {0.0000} -0.1041 {0.0000} 0.4392 {0.0000}

-6.9033 {0.0000} -0.6545 {0.0000} -0.2785 {0.0000} 0.3474 {0.0000}

***

-0.0740 {0.0000} 0.9821 {0.0000} 0.0152 {0.0205} 0.0723 {0.0000}

***

-0.1272 {0.0000} 0.9609 {0.0000} 0.0150 {0.0590} 0.0841 {0.0000}

***

-0.5074 {0.0000} 0.9040 {0.0000} -0.0020 {0.8912} 0.3433 {0.0000}

***

-0.1719 {0.0000} 0.9548 {0.0000} -0.0060 {0.5975} 0.1353 {0.0000}

***

LogL Q(20)
*

-2058 22.6680 {0.3054} 5.7428 {0.9992} 9.2697 {0.9796}

2036 25.7302 {0.1749}

-2061 21.0281 {0.3955}

-1102 28.5201 {0.0976}

IDRUSD H1 H2 H0 Conditional mean equation -0.0042 -0.0241 *** 0.0003 {0.5201} {0.0002} {0.9790} 0.0648 ** 0.0857 *** -0.0489 {0.0165} {0.0020} {0.0187} Conditional variance equation -0.2944 *** -0.4704 *** -0.0730 {0.0000} {0.0000} {0.0000} 0.9331 *** 0.8642 *** 0.9827 {0.0000} {0.0000} {0.0000} -0.0735 *** -0.0438 *** 0.0148 {0.0000} {0.0000} {0.2301} 0.1879 *** 0.4652 *** 0.0725 {0.0000} {0.0000} {0.0000} Estimation diagnostics -96 -1112 -1922 193.0183 *** 52.3700 *** 15.5029 {0.0000} {0.0001} {0.7469} 1287 48.5800 {0.0004} 6.6898 {0.9976} 18.6775 {0.5429} 35.6763 {0.0168}
***

-1894 12.2016 {0.9090}

**

-789 14.4769 {0.8055} 15.8247 {0.7274} 20.6516 {0.4179}

44 88.4089 {0.0000} 6.8849 {0.9970}

***

-833 33.3299 {0.0310} 29.5565 {0.0774}

**

Q (20)

22.0348 {0.3386}

85.3750 {0.0000}

***

21.9374 {0.3439}

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively. Q(20) and Q2(20) are Ljung-Box portmanteau test of white noise for linear and non-linear (squared) standardized residual. 142

Table 6.2 - Continued

H2
* ***

H0
***

NZDUSD H1 H0 H2

TWDUSD H1

c 0.0263 {0.0876} -0.0002 {0.9916}

*** *** *** *** *** *** ***

**

Lag, 1

0.0190 {0.1285} 0.0004 {0.9840}

***

-0.0061 {0.0038} 0.0118 {0.3160}

0.0031 {0.3981} -0.0906 {0.0005}

0.0061 {0.0000} 0.2263 {0.0000}

0.0084 {0.0400} 0.2352 {0.0000}

***

c
*** *** *** ***

***

***

h
*** ***

***

***

***

1
*** *** ***

**

***

***

**

-0.0743 {0.0003} 0.9772 {0.0000} -0.0261 {0.0422} 0.0757 {0.0001}

***

***

-2.7640 {0.0000} 0.3248 {0.0000} 0.0189 {0.2708} -0.2753 {0.0000}

-0.0785 {0.0000} 0.9753 {0.0000} -0.0254 {0.0022} 0.0795 {0.0000}

-0.4344 {0.0000} 0.9029 {0.0000} 0.0403 {0.0000} 0.2557 {0.0000}

-0.3115 {0.0000} 0.9346 {0.0000} -0.0480 {0.0000} 0.1524 {0.0000}

***

-0.2169 {0.0000} 0.9530 {0.0000} 0.0094 {0.0438} 0.1296 {0.0000}

***

LogL Q(20)
***

-2259 18.1102 {0.5802}

1546 20.2831 {0.4403}

-2249 19.8618 {0.4666}

PHPUSD H0 H1 H2 Conditional mean equation -0.0179 *** 0.0000 0.0000 {0.0003} {1.0000} {1.0000} -0.1576 *** 0.1970 *** 0.0828 {0.0000} {0.0000} {0.0000} Conditional variance equation -0.2935 *** -0.3337 *** -0.0131 {0.0000} {0.0000} {0.0000} 0.9661 *** 0.8929 *** 0.9995 {0.0000} {0.0000} {0.0000} -0.0566 *** 0.1659 *** -0.0253 {0.0000} {0.0000} {0.0000} 0.3189 *** 0.2792 *** 0.0177 {0.0000} {0.0000} {0.0000} Estimation diagnostics -686 -791 -1119 ** 13.6570 35.2676 33.0935 {0.0330} {0.8475} {0.0187}
**

-38 33.5468 {0.0294}

**

456 320.47928*** p {0.0000}

-97 174.3883 {0.0000}

***

17.1952 326.1031 15.4165 14.8839 0.1044 1.7050 15.9195 1.2994 8.9048 {0.6403} {0.0000} {0.7521} {0.7830} {1.0000} {1.0000} {0.7216} {1.0000} {0.9840} Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively. Q(20) and Q2(20) are Ljung-Box portmanteau test of white noise for linear and non-linear (squared) standardized residual.

Q2(20)

143

Table 6.2 - Continued

Panel B. Asia-Pacific currencies/EUR AUDEUR H0 H1 H0

***

H2

JPYEUR H1 H2 H0

KRWEUR H1

H2

c
***

Lag, 1

0.0016 {0.9034} 0.0304 {0.1129}

*** ** *** *** *** *** *** *** ***

-0.0006 {0.7532} 0.0371 {0.0006}

***

-0.0004 {0.9673} 0.0499 {0.0067}

-0.0339 {0.1152} -0.1073 {0.0011}

-0.0029 {0.0006} 0.0747 {0.0456}

-0.0190 {0.1107} 0.0076 {0.6870}

-0.0084 {0.6689} -0.2128 {0.0000}

0.0029 {0.4494} 0.0082 {0.7478}

***

-0.0101 {0.5245} -0.2168 {0.0000}

***

c
*** *** *** *** ***

***

***

***

h
*** ***

***

***

***

***

**

1
*** *** *** ***

**

***

***

-0.0506 {0.0000} 0.9972 {0.0000} -0.0062 {0.2382} 0.0632 {0.0000}

***

***

-0.0176 {0.0000} 0.9989 {0.0000} -0.0007 {0.5389} 0.0201 {0.0000}

***

-0.0590 {0.0000} 0.9950 {0.0000} -0.0146 {0.1741} 0.0710 {0.0000}

-0.1242 {0.0000} 0.9549 {0.0000} -0.0417 {0.0014} 0.1254 {0.0000}

-0.0302 {0.0000} 0.9981 {0.0000} 0.0249 {0.0000} 0.0427 {0.0000}

-0.1041 {0.0000} 0.9898 {0.0000} 0.0126 {0.0733} 0.1230 {0.0000}

-1.3028 {0.0000} -0.5282 {0.0000} 0.0902 {0.0114} 0.0915 {0.0063}

***

-3.9779 {0.0000} 0.0336 {0.0000} 0.1525 {0.0000} -0.0896 {0.0000}

***

-1.2171 {0.0000} -0.4126 {0.0488} 0.1010 {0.0028} 0.1208 {0.0856}

LogL Q(20) 26.3950 {0.1532} 4.7855 {0.9998} 19.1328 {0.5132} 26.7730 {0.1418}

-1970 22.6122 {0.3082}

1409 19.1871 {0.5097}

-1942 22.1041 {0.3349}

-1262 23.5135 {0.2643}

IDREUR H1 H2 H0 Conditional mean equation -0.0030 -0.0163 -0.0198 {0.2498} {0.4156} {0.1382} 0.0768 *** -0.0689 ** -0.0112 {0.0001} {0.0293} {0.6107} Conditional variance equation -0.5769 *** -0.1996 *** -0.1018 {0.0000} {0.0000} {0.0000} 0.9020 *** 0.8977 *** 0.9915 {0.0000} {0.0000} {0.0000} -0.0184 *** -0.0518 *** 0.0130 {0.0080} {0.0012} {0.1163} 0.3997 *** 0.1817 *** 0.1214 {0.0000} {0.0000} {0.0000} Estimation diagnostics 426 -1293 -2044 26.9773 19.8119 34.2313 ** {0.0246} {0.1359} {0.4698} 1201 34.8739 {0.0208} 11.4433 {0.9339}
**

-2026 19.9708 {0.4598} 25.5297 {0.1819}

-1161 25.1487 {0.1958} 77.1249 {0.0000}

***

753 26.0311 {0.1648} 3.6408 {1.0000}

-1167 21.8406 {0.3492} 55.2080 {0.0000}

***

Q (20)

20.7388 {0.4127}

12.4382 {0.9002}

19.8339 {0.4684}

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively. Q(20) and Q2(20) are Ljung-Box portmanteau test of white noise for linear and non-linear (squared) standardized residual.

144

Table 6.2 - Continued

H2
**

H0
***

NZDEUR H1 H0 H2

TWDEUR H1

c
*

Lag, 1

0.0108 {0.4962} -0.0107 {0.5844}

* *** *** *** *** *** *** ***

0.0169 {0.0000} -0.0161 {0.0558}

0.0015 {0.9236} 0.0232 {0.1985}

-0.0168 {0.2141} -0.1530 {0.0000}

0.0004 {0.4440} 0.0372 {0.0047}

-0.0025 {0.8581} -0.1238 {0.0000}

***

c
*** *** *** ***

***

***

h
***

***

***

***

1
*** ***

***

-0.0967 {0.0000} 0.9836 {0.0000} -0.0063 {0.3545} 0.1105 {0.0000}

***

***

-6.3101 {0.0000} -0.6682 {0.0000} -0.1497 {0.0000} 0.3206 {0.0000}

-0.0866 {0.0000} 0.9856 {0.0000} -0.0071 {0.2886} 0.0985 {0.0000}

-0.0320 {0.0000} 0.9969 {0.0000} 0.0052 {0.2822} 0.0375 {0.0000}

-1.5399 {0.0000} 0.5859 {0.0000} -0.2126 {0.0000} 0.4720 {0.0000}

***

-0.0288 {0.0000} 0.9961 {0.0000} 0.0037 {0.2168} 0.0328 {0.0000}

***

LogL Q(20)
* **

-2194 22.6764 {0.3050}

860 30.4885 {0.0623}

-2169 19.6723 {0.4786}

PHPEUR H0 H1 H2 Conditional mean equation -0.0298 * -0.0067 -0.0470 {0.0911} {0.5557} {0.0169} -0.1246 *** -0.0037 -0.0551 {0.0000} {0.8950} {0.0596} Conditional variance equation -1.2535 *** -0.0714 * -1.1350 {0.0000} {0.0919} {0.0000} -0.5233 * 0.9836 *** -0.3412 {0.0826} {0.0000} {0.0000} 0.0550 -0.0011 0.0326 {0.4460} {0.9765} {0.3482} 0.1884 *** 0.0784 ** 0.0193 {0.0000} {0.0457} {0.5834} Estimation diagnostics -1313 -599 -1247 ** 34.5118 18.9837 31.5698 {0.0229} {0.5229} {0.0481}
**

-1858 15.2576 {0.7615}

372 69.0554 {0.0000}

***

-1872 13.2197 {0.8678}

24.5036 18.2001 23.7716 36.9992 3.4944 19.2829 18.5822 1.2376 12.6867 {0.2211} {0.5742} {0.2525} {0.0117} {1.0000} {0.5035} {0.5491} {1.0000} {0.8904} Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively. Q(20) and Q2(20) are Ljung-Box portmanteau test of white noise for linear and non-linear (squared) standardized residual.

Q2(20)

145

6.4.2. The Feds target interest rate news spillover effects Table 6.3 shows the estimates for the spillover effects of the Feds and the ECBs target rate surprises as modeled in (6.2a) and (6.2b). Three currencies (AUD, NZD and TWD) depreciated significantly against the USD over H0, which was mostly coming from significant response in H2. On the other hand, three others (IDR, JPY, and KRW) appreciated mostly during H1. It can be conjectured that an unexpected hike in Feds target rate could be viewed as a strong commitment to combat inflation. As a result, such unexpected tightening may lead to a depreciation of the USD. However, such effect is not persistent evidenced by the depreciation of the Yen in the following day. There is some evidence of slow absorption of Fed news as the news effect is either significant in H2 only or significant exchange rate response in H2 is persisting into the longer horizon of H1. There is also evidence for the Feds spillover impacts on the conditional variance of the exchange rates changes. In general, a volatility reducing influence of the Feds news has been found in four currencies in seven cases during at least one time horizon. The volatility raising effect has been shown in two cases of IDR/USD and JPYUSD pairs. However, this effect does not persist to the daily horizon H1. It can be conjectured that, the Asia Pacific markets participants might be at an information disadvantage regarding forming expectations on the US Feds interest rate decisions compared to the US market participants. When these markets open for trading, clearer US market movements would have been established and the potential implications of the Fed surprises on the other markets would have also been well understood. Therefore, a lower level of diversity of opinions upon the spillover effect
146

of the US announcements on the other markets would have been shown. 49 Some evidence for the delayed volatility response has been found where volatility effect is significant only in intradaily horizon for two cases (IDR, KRW), and in all three horizons for the other two (JPY and TWD). Table 6.3 - The Feds and the ECBs target interest rate news spillover effects
This table reports the estimation results of the EGARCH models described in (6.2a) and (6.2b) for daily changes in Asia Pacific exchange rates against the USD and the EUR against the Feds and the ECBs news, respectively. H0 is daily horizon over the close on calendar day t-1 to the close on calendar day t; H1 is overnight horizon over the close of day t-1 to the open on day t; and H2 is intraday horizon measured over the open on day t to the close on day t. P-values are in braces.

y t = [RHS of (6.1a)] + News News t 1

(6.2a) (6.2b)

ln ht = [RHS of (6.1b)] + News Newst 1 + MacroAnn,i MacroAnni ,t 1

i

Panel A. The Feds target interest rate news spillover effects

Exchange rate AUDUSD IDRUSD JPYUSD KRWUSD NZDUSD PHPUSD TWDUSD

H0 -0.8224 {0.0000} 2.3544 {0.0459} 0.1735 {0.6197} -0.0396 {0.8795} -1.0199 {0.0000} -0.0139 {0.9264} -0.0720 {0.0000}

***

**

***

***

The Feds' target interest rate news spillover effects FedNews FedNews H1 H2 H0 H1 -0.0008 -0.8000 *** -0.0675 -1.7335 {0.7776} {0.0000} {0.6678} {0.3522} 0.0168 *** 1.2239 *** 3.8371 0.6724 {0.0000} {0.0072} {0.3572} {0.6180} 0.0629 *** -0.2939 *** -0.2986 *** -2.3214 *** {0.0000} {0.0054} {0.0039} {0.0018} 0.3831 * -0.3733 -0.4800 -0.0063 {0.0820} {0.1321} {0.1492} {0.9905} -0.0021 ** -1.0091 *** -0.0894 -5.5684 *** {0.0191} {0.0000} {0.4937} {0.0000} 0.0166 -0.2322 0.3401 0.5052 {0.9217} {0.2365} {0.3175} {0.2025} 0.2841 -0.1349 *** -0.2991 * -1.4356 ** {0.2631} {0.0058} {0.0526} {0.0263}

H2 -0.0899 {0.5814} 2.7223 {0.0006} 0.6677 {0.0054} -0.7658 {0.0654} -0.0665 {0.6234} 0.4727 {0.4528} -0.5827 {0.0491}

***

***

**

49

For brevity, estimations results of the macroeconomic announcement dummies are not reported.

Instead, the main results are summarized here. The macroeconomic announcement dummies have significant effects on the conditional volatility of currencies pairs. While the GDP and Trade deficit (TD) announcements tend to reduce market volatility, the reverse is observed in cases of the CPI and Unemployment (UE) announcements. Retail sales growth rate (RET) has unclear impact on the volatility with four cases each with higher and lower volatility.

147

Table 6.3 Continued

Panel B. The ECB' target interest rate news spillover effects The ECB' target interest rate news spillover effects Exchange ECBNews ECBNews rate H0 H1 H2 H0 H1 AUDEUR 1.6680 *** 0.0782 2.0502 -0.8884 -7.4612 *** {0.0000} {0.7161} {0.3692} {0.5638} {0.0069} IDREUR -7.5702 *** -0.8861 -7.8851 * 2.2473 10.5430 {0.0000} {0.2218} {0.0667} {0.5948} {0.4577} JPYEUR 1.3821 0.0238 1.1181 0.4231 -4.2560 {0.4337} {0.8925} {0.4755} {0.8592} {0.5489} KRWEUR -10.3437 ** 0.1816 -9.9304 *** 6.9093 2.0978 *** {0.0457} {0.4368} {0.0025} {0.2882} {0.0007} NZDEUR 1.7187 -0.0367 2.4347 -0.5149 -2.7469 {0.3520} {0.9433} {0.9804} {0.8542} {0.5157} PHPEUR -4.6127 1.0884 -7.9111 *** 7.6355 -3.4551 {0.2848} {0.3973} {0.0027} {0.2499} {0.3148} TWDEUR -3.2432 ** 0.0166 -3.5016 ** -1.4193 *** -4.8834 *** {0.0203} {0.9076} {0.0213} {0.0000} {0.0002} Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

H2 -1.0091 {0.5985} 3.3550 {0.1258} 1.1037 {0.6335} 2.7455 {0.8297} 0.0568 {0.9804} -3.1838 {0.1844} -2.1472 {0.0003}

***

6.4.3. The ECBs target interest rate news spillover effects In general, similar to the Feds news, the ECBs news also shows tardy impact on the Asia Pacific currencies with three currencies (IDR, KRW and TWD) depreciate against the EUR in the intraday and daily horizons while the PHP depreciates during the intraday horizon only. On the conditional volatility, some evidence for a volatility reducing effect of the ECBs news in two exchange rates (AUD and TWD) has been found. However, it is also observed that the ECBs news has volatility increasing effect on the conditional volatility of the KRW rate during overnight only. It is conjectured that even though the ECBs news has injected some insights about the euro areas economies that might help to resolve uncertainties and heterogeneities in beliefs among market participants, such temporary homogeneity may eventually evaporate as the market participants may be still uncertain about the future course of the ECBs target rate. In general, the ECBs news has been absorbed relatively quickly with

148

some currency pairs responded quickly as evidenced by significant news coefficient in overnight horizon only (AUD/EUR and KRW/EUR), while in the other (TWD/EUR) the news effect shows up in all three horizons.50 The results tend to suggest that the impacts of the ECBs news are clearer than the Feds news in the Asia Pacific markets. This may be due to the timing effect where the ECBs news tends to confirm the earlier arrival of the Feds news. This hypothesis is empirically tested in Section 6.5.1. 6.5. Robustness checks 6.5.1. Joint spillover effects of the target interest rate news To the extent that the Feds rate decisions can potentially influence those of the ECB due to the earlier arrival and the role that the US economy plays in the world economy, it is possible that the impacts of the ECBs rate decisions on the Asia Pacifics currencies might be weaker or even insignificant as the market participants might regard the ECB news as less newsworthy. As shown in Chapter 5, Granger causality tests revealed that there is a strong evidence of the Feds FOMC Granger causing the ECBs Governing Council in making interest rate decisions, while the evidence for the reverse is weak. Thus, it is necessary to control for the other central banks news effect to check for the robustness of the target rate news impacts and to further investigate the relationship between the two sets of news.

50

The euro areas macroeconomic announcement dummies have significant effects on the conditional

volatility where the CPI tends to increase market volatility; the other announcements tend to reduce volatility in most of currency pairs at least in one time horizon.

149

Equations (6.3a) and (6.3b) include both the Feds and ECBs target news coefficients and both the USs and the euro areas macroeconomic announcements dummies.
y t = [RHS of (6.1a)] + FedNews FedNews t 1 + ECBNews ECBNews t 1

(6.3a)

ln h t = [RHS of (6.1b)] + FedNews FedNews + ECBNews ECBNews

t 1

t 1

+ USMacroAnn ,iUSMacroAnn
i

i , t 1

(6.3b)

+ EuroareaMa
i

croAnn , i

EuroareaMa croAnn

i , t 1

The ECBs news coefficient on the mean would be insignificant if the ECBs news has no impact on the Asia Pacifics currencies if the ECBs news is stale when it hits the Asia Pacific. However, if the ECBs news has its own merits, the ECBs news coefficient is expected to be significant. On the variance equation, the impact of the ECBs news would once again depends on its supplemental role for the Feds news in either raising or lowering the level of market uncertainty on the exchange rates of the Asia Pacifics currencies against the USD and the EUR. The estimates for the joint spillover effects of the Feds and the ECBs target rate news as modeled in (6.3a) and (6.3b) are reported in Table 6.4. Overall, the results suggest that the separate effects of the Feds and the ECBs news as reported in the previous sections are robust. For the exchange rates against the USD, while the Feds effect is robust, the ECBs news also has significant impacts on the exchange rates against the USD. However, the ECBs effect on exchange rates against the USD is mixed with two cases where the KRW and the PHP depreciates against the USD (mostly overnight), and three other cases where the IDR, PHP and the TWD appreciates against the USD (mostly during the intraday horizon). It is conjectured that the late arrival of the ECBs news might help to confirm the prevailing belief

150

that the Fed is going to take further steps in fighting against inflation. On the volatility, in response to unexpected hikes in the ECBs target rate, four currencies show lower volatility in at least one time horizon, while the other two (JPY and TWD) show higher volatility overnight only. This finding reaffirms the hypothesis for the supplementary role that the ECBs news has played to the Feds news in Asia Pacific currencies against the USD. For the exchange rates against the EUR, while the ECBs effect is marginally stronger, the Feds news tends to lead to an appreciation of the Asia Pacific currencies against the EUR mostly during the overnight horizon. On the other hand, the Feds news seems to show a volatility reducing effect across three horizons. Theses findings suggest that the earlier arrival of the Feds news might send a leading signal to the market that the ECB might also increase its target rate to combat against the inflation. Thus, this signaling effect might result in a depreciation of the EUR and a lower volatility. 6.5.2. Asymmetric effects There is evidence for an asymmetric influence of the past innovations as shown by a negative 1 in most of the cases for the daily and overnight horizons respectively. The potential of asymmetric influences of unexpected rises and unexpected fall of the Feds and the ECBs target rate announcements is investigated by equations (6.4a) and (6.4b). The overall news variable is disaggregated into two types of news, unexpected rises and unexpected cuts. As suggested in the literature, it is expected that unexpected rate rises would have a larger impact on the Asia Pacific currency markets than unexpected rate cuts. This is empirically modeled by partitioning the Newst-1 variable into a rate rise surprise component, News_Rt-1, and a

151

rate fall, News_Ft-1, and add these to equations (6.1a) and (6.1b) as below. The potential for an asymmetric effect is ascertained by examining the difference between the two coefficients in terms of their signs and magnitudes.
y t = [RHS of (6.1a)] + News _ R News _ Rt 1 + News _ F News _ Ft 1

(6.4a) (6.4b)

ln ht = [RHS of (6.1b)] + News _ R News _ Rt 1 + News _ F News _ Ft 1 +

MacroAnn,i

MacroAnni ,t 1

Table 6.5 shows the estimates of the asymmetric responses of the conditional mean and variance to positive and negative surprises. Weak evidence has been found for the asymmetric impact of the Feds and the ECBs rate news. On the conditional mean, a marginal equal number of responses to the Feds and ECBs unexpected rate rises and rates cuts was observed. However, while unexpected rate rises led to more appreciations than depreciations, the opposite is true for rate cuts. Furthermore, only KRW/USD and PHP/EUR show asymmetric effects during H0 and H2 where KRW and PHP appreciate against the USD and EUR respectively in response to unexpected rate increases and depreciate in response to unexpected rate decreases. Both the Feds and ECBs unexpected rate increases have unclear impact on the conditional volatility while unexpected rate decreases led to a lower volatility in most of the cases. Furthermore, as far as magnitude is concerned, unexpected rate increases tend to have a larger impact than unexpected rate decreases. Furthermore, in most cases, the hypothesis that the magnitudes of positive and negative news coefficients are equal in the variance equation is rejected.

152

Table 6.4 - Joint spillover effects of the US Feds and the ECBs target interest rate news

This table reports the estimation results of the EGARCH models described in (6.3a) and (6.3b) for daily changes in Asia Pacific exchange rates against the USD and the ECB against the overall surprises of the Fed and the ECB. H0 is daily horizon over the close on calendar day t-1 to the close on calendar day t; H1 is overnight horizon over the close of day t-1 to the open on day t; and H2 is intraday horizon measured over the open on day t to the close on day t. P-values are in braces.

y t = [RHS of (6.1a)] + FedNews FedNews t 1 + ECBNews ECBNews t 1

i i

(6.3a) (6.3b)

ln ht = [RHS of (6.1b)] + FedNews FedNewst 1 + USMacroAnn,iUSMacroAnni ,t 1 + ECBNews ECBNewst 1 + EuroareaMacroAnn,i EuroareaMacroAnni ,t 1

Conditional variance equation

Panel A. Asia-Pacific currencies/USD

Exchange rate
***

Conditional mean equation

AUDUSD
** *** ***

***

**

IDRUSD

***

***

***

JPYUSD
*

***

**

KRWUSD
***

**

NZDUSD
***

***

***

***

PHPUSD
** ***

***

***

***

***

TWDUSD

H0 -0.8601 {0.0000} 2.3321 {0.0055} 0.1710 {0.5043} -0.1309 {0.6560} -1.0305 {0.0000} -0.0088 {0.9518} -0.0706 {0.0000}

***

FedNews H1 -0.0008 {0.6097} 0.6056 {0.0002} 0.0467 {0.2834} 0.4427 {0.0595} -0.0023 {0.3718} 0.0800 {0.2630} 0.1441 {0.2323} H0 1.1685 {0.5301} 1.0519 {0.2139} -0.8228 {0.5935} -1.8519 {0.0784} 0.6900 {0.6894} -0.6677 {0.3729} 0.0063 {0.0000} ECBNews H1 -0.0008 {0.8130} 0.6056 {0.2060} 0.0646 {0.9300} -0.9056 ** {0.0236} 0.0052 {0.8943} -0.6508 *** {0.0001} 2.0736 ** {0.0499} H2 1.3215 {0.5213} 2.1164 {0.0027} -0.3150 {0.8416} 1.0643 {0.3692} 0.9536 {0.6335} 0.9527 {0.0003} 0.0785 {0.2649} H0 -0.1740 {0.1853} 3.7294 {0.0001} 0.1958 {0.1055} -0.2943 {0.5539} -0.1622 {0.3384} -0.4292 {0.2136} -0.3798 {0.3953} FedNews H1 -3.7846 {0.2228} 1.8352 {0.2320} -2.6698 {0.0001} -0.1336 {0.7700} -4.9186 {0.0000} -0.6245 {0.0000} 0.9066 {0.1698}

H2 -0.8401 {0.0000} 1.2610 {0.0124} 0.1349 {0.6928} -0.4294 {0.1477} -1.0319 {0.0000} -0.1802 {0.2682} -0.1381 {0.0312}

H2 -0.2406 {0.0492} 2.8938 {0.0022} 0.2071 {0.9712} -0.5477 {0.0758} -0.1437 {0.3514} 0.1036 {0.8245} -0.5865 {0.1132}

H0 0.9493 {0.4835} -9.5531 {0.0532} 0.4754 {0.6692} 3.1051 {0.3875} 0.3770 {0.7925} -12.2305 {0.0056} -4.4152 {0.4281}

ECBNews H1 -11.3158 {0.1172} 6.6225 {0.7269} 25.3592 {0.0413} -2.6289 {0.0450} -13.1376 {0.0000} -14.0620 {0.0014} 9.9415 {0.0683}

H2 1.6823 {0.2166} -2.7467 {0.6206} -0.1971 {0.7674} 2.8674 {0.0772} 0.6982 {0.6712} -9.5110 {0.0016} -9.1153 {0.5105}

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

153

Table 6.4 Continued

Conditional variance equation

Panel B. Asia-Pacific currencies/EUR

Exchange rate
*** ***

Conditional mean equation

AUDEUR
*** * ** **

**

***

IDREUR
** *

**

**

**

JPYEUR
*** ***

**

**

KRWEUR
*** ***

***

NZDEUR
**

**

***

PHPEUR
*** *** ** ***

TWDEUR

H0 0.0663 {0.7315} -0.9813 {0.0202} 0.6287 {0.0115} 0.5451 {0.4841} 0.1016 {0.5703} -0.1313 {0.7609} 0.8338 {0.0003}
***

***

FedNews H1 0.0554 {0.0299} 0.2414 {0.0111} 0.0605 {0.0239} 0.0385 {0.8446} 0.0112 {0.7496} -0.1456 {0.2317} 0.0480 {0.0000} H0 1.3049 {0.5732} -7.3882 {0.0813} 0.2513 {0.8971} -10.7703 {0.0001} 1.3380 {0.1349} -5.3305 {0.2342} -2.7775 {0.0003} ECBNews H1 -0.0129 {0.9436} -0.8805 {0.1495} -0.0745 {0.6295} 0.1087 ** {0.0284} -0.2095 {0.3754} 1.1289 {0.2828} 0.0097 {0.9470} H2 1.7147 {0.3922} -7.4357 {0.0402} -0.2389 {0.9003} -10.2564 {0.0000} 1.8911 {0.3754} -6.8656 {0.0379} -3.9196 {0.0281} H0 -0.7429 {0.0000} -0.8221 {0.1975} 0.2344 {0.0730} -0.1675 {0.5451} -0.4346 {0.0044} 0.5315 {0.4355} -0.2741 {0.0000} H2 -0.7510 {0.0000} -0.8423 {0.1577} 0.1539 {0.2573} 1.9419 {0.0627} -0.3939 {0.0207} -0.7888 {0.1118} -0.2471 {0.0000} H0 1.4678 {0.4432} 3.0124 {0.6173} -1.1978 {0.6243} 0.1433 {0.9493} 0.6636 {0.5646} 6.5865 {0.3281} -0.3059 {0.0000} FedNews H1 -0.1000 {0.8699} -0.9284 {0.0119} -0.0582 {0.9536} 4.6687 {0.1096} -2.8266 {0.0000} 0.1755 {0.5805} -0.4612 {0.2938}

***

H2 0.0051 {0.9782} -1.1582 {0.0037} 0.6375 {0.0263} 0.1220 {0.8856} 0.0721 {0.7496} -0.0077 {0.9417} 0.8382 {0.0006}

***

ECBNews H1 -8.5265 {0.0059} 8.9643 {0.0141} 2.7098 {0.7249} -41.8308 {0.0001} 11.8021 {0.0000} -3.2561 {0.4966} -9.3456 {0.0180}

**

H2 1.2015 {0.4828} 3.3971 {0.5895} 0.0423 {0.9831} 3.6371 {0.7821} 1.1695 {0.5638} 4.4074 {0.5621} -0.0237 {0.0000}

***

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

154

Table 6.5 - Asymmetric news spillover effects

This table reports the estimation results of the EGARCH models described in (6.4a) and (6.4b) for daily changes in Asia Pacific exchange rates against the USD and the EUR against the Feds and the ECBs news, respectively. H0 is daily horizon over the close on calendar day t-1 to the close on calendar day t; H1 is overnight horizon over the close of day t-1 to the open on day t; and H2 is intraday horizon measured over the open on day t to the close on day t. P-values are in braces.

y t = [RHS of (6.1a)] + News _ R News _ Rt 1 + News _ F News _ Ft 1

i

(6.4a) (6.4b)

ln ht = [RHS of (6.1b)] + News _ R News _ Rt 1 + News _ F News _ Ft 1 + MacroAnn,i MacroAnni ,t 1

***

***

Panel A. Asia-Pacific currencies/USD AUDUSD H0 H1 -1.7902 ** -0.1005 FedNews_R {0.0239} {0.6957} -0.7350 *** -0.0010 FedNews_F {0.0000} {0.9652} -1.3535 ** 2.1698 * FedNews_R {0.0432} {0.0603} -0.0472 -2.3254 *** FedNews_F {0.7458} {0.0000} H0 3.2816 {0.0000} -0.4352 {0.4819} 7.2791 {0.0000} 0.6887 {0.4005}
*** ***

H2 -1.6461 {0.1151} -0.7229 {0.0000} -1.1451 {0.1904} -0.0706 {0.6563}

H2 -0.8912 {0.4518} -0.1803 {0.4872} 2.9695 {0.2431} 0.5793 ** {0.0290}

H0 1.5274 *** {0.0019} -0.4395 *** {0.0000} -1.6727 {0.1416} -0.3921 {0.2368}

KRWUSD H1 0.2117 {0.5179} 0.3844 {0.1333} 1.8215 {0.5205} -0.2479 {0.6173}

H2 0.9189 ** {0.0463} -0.8918 *** {0.0000} -2.9150 *** {0.0000} -0.8539 *** {0.0000}

Mean Eq
***

Var Eq

1.6279 {0.2020} 4.0701 {0.0436}

**

0.1488 {0.6997} 13.9323 {0.0002}

***

0.7635 {0.3822} 1.5407 {0.2145}

15.3226 {0.0001} 17.1033 {0.0000}

IDRUSD JPYUSD H1 H2 H0 H1 0.5558 * 1.9500 *** -0.7059 0.0264 *** {0.0880} {0.0000} {0.3849} {0.0000} 0.0152 -0.2888 0.2491 0.0613 *** {0.9788} {0.7560} {0.4996} {0.0000} 1.9351 *** 5.2123 *** 0.2211 -18.4910 *** {0.0000} {0.0001} {0.7624} {0.0000} 0.0779 0.8640 0.2900 *** -1.9808 *** {0.8043} {0.4916} {0.0031} {0.0021} Tests of equality between FedNews_R and FedNews_F 0.6761 7.6204 *** 1.1958 5.7843 ** {0.4109} {0.0058} {0.2742} {0.0162} 10.6792 *** 5.0863 ** 0.0082 57.6797 *** {0.0011} {0.0241} {0.9280} {0.0000}

0.3281 {0.5668} 0.8692 {0.3512}

15.5295 *** {0.0001} 1.2781 {0.2582}

0.1727 {0.6778} 0.5051 {0.4773}

3.0875 *** {0.0000} 32.7038 *** {0.0000}

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

155

Table 6.5 Continued

FedNews_R

***

FedNews_F

***

***

FedNews_R
*

FedNews_F

H0 -3.2806 {0.0009} -0.7228 {0.0000} -0.3938 {0.6291} -0.0865 {0.5175}

NZDUSD H1 -0.0592 ** {0.0349} -0.0021 ** {0.0349} -8.4922 *** {0.0005} -5.4732 *** {0.0000} H0 0.3870 {0.2084} -0.0720 {0.2572} -0.1403 {0.8962} -0.3385 {0.0782}

TWDUSD H1 0.0633 {0.7894} 0.0195 {0.6617} 0.8603 {0.2929} -0.8542 *** {0.0000}

H2 0.2286 {0.6648} -0.1400 {0.0001} 1.2327 {0.3585} -0.6906 {0.0434}

**

Mean Eq

**

Var Eq

6.6137 {0.0101} 0.1506 {0.6980}

0.7986 {0.3715} 1.4567 {0.2275}

PHPUSD H2 H0 H1 H2 -3.1836 *** 0.6915 0.3319 *** 0.3161 {0.0007} {0.3686} {0.0008} {0.6657} -0.7031 *** -0.0911 0.0381 -0.2981 ** {0.0000} {0.2742} {0.6687} {0.0164} -0.1193 3.9385 *** 1.8257 0.5630 {0.8617} {0.0014} {0.4328} {0.8863} -0.0668 -0.8656 * -0.0447 0.4542 {0.5889} {0.0675} {0.9506} {0.5542} Tests of equality between FedNews_R and FedNews_F 6.8591 *** 1.0251 6.8310 *** 0.7010 {0.0088} {0.3113} {0.0090} {0.4025} 0.5707 0.0010 0.0060 12.9869 *** {0.9384} {0.0003} {0.4500} {0.9750} 2.1359 {0.1439} 0.0330 {0.8558} 0.0332 {0.8555} 4.3519 {0.0370}
**

0.4842 {0.4865} 1.7756 {0.1827}

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

156

Table 6.5 Continued

Panel B. Asia-Pacific currencies/EUR AUDEUR H0 H1 7.4145 * 0.6616 ECBNews_R {0.0516} {0.1580} -0.5430 0.0180 ECBNews_F {0.7877} {0.9124} 1.6895 -2.0867 ECBNews_R {0.6125} {0.7653} -2.0229 -9.5168 ** ECBNews_F {0.2492} {0.0163}
*

H2 6.5542 {0.0898} 0.0553 {0.9759} 1.7390 {0.3580} -2.1982 {0.0560}

H0 5.6240 {0.7163} -10.9583 {0.0009} 26.8006 {0.0410} -2.3592 {0.6163}

H2 2.5016 {0.5732} -0.2642 {0.9239} 1.3352 {0.5227} 1.0945 {0.4128}

H0 -12.6814 {0.6072} -10.4505 *** {0.0006} 18.8404 * {0.0620} 0.5070 {0.8822}

KRWEUR H1 0.4024 {0.3981} -0.0048 {0.9764} 11.1931 {0.1479} 1.1134 {0.7557}

H2 -12.6107 {0.2569} -10.0878 *** {0.0000} 19.4845 *** {0.0005} 1.0218 {0.7444}

Mean Eq
**

Var Eq

3.4062 {0.0650} 1.3224 {0.2502}

1.7396 {0.1872} 0.7820 {0.3765}

2.6597 {0.1029} 4.4612 {0.0347}

1.0286 {0.3105} 4.4216 {0.0355}

IDREUR JPYEUR H1 H2 H0 H1 -1.1306 3.8941 3.1723 0.0115 {0.6163} {0.8124} {0.3119} {0.9601} *** -0.5920 -11.0158 *** 0.2566 -0.0596 {0.5827} {0.0014} {0.9674} {0.9601} ** -13.9507 29.2105 ** 0.1543 -23.0514 *** {0.8468} {0.0381} {0.9674} {0.0001} 13.4142 -0.8637 0.1543 -6.6181 * {0.3100} {0.8609} {0.8212} {0.0878} Tests of equality between ECBNews_R and ECBNews_F 0.0704 0.8130 0.5038 0.0695 ** {0.7908} {0.3672} {0.4779} {0.0194} ** 0.1134 4.2645 ** 0.0116 5.4607 ** {0.7363} {0.0389} {0.9142} {0.0194} 0.2829 {0.5948} 0.0109 {0.9170} 0.0083 {0.9272} 4.4637 ** {0.0346}

0.5596 {0.4544} 1.2728 {0.2592}

0.0495 {0.8239} 12.0605 *** {0.0005}

ECBNews_R

ECBNews_F

ECBNews_R

ECBNews_F

H0 1.5521 {0.6395} 1.7762 {0.4689} -1.0570 {0.8071} -0.3262 {0.8961}

NZDEUR H1 0.2416 {0.7014} -0.2767 {0.7344} -12.6745 ** {0.0180} 3.5336 {0.4604} H0 0.2382 {0.9238} -4.6116 ** {0.0139} 0.0192 {0.9887} -1.7897 *** {0.0001}

TWDEUR H1 -2.3913 *** {0.0000} 0.1821 {0.2054} 11.6450 *** {0.0000} -5.5029 *** {0.0060}

H2 0.5650 {0.8723} -5.2833 ** {0.0241} 1.2424 {0.1204} -2.0739 *** {0.0000}

Mean Eq

Var Eq

0.0036 {0.9522} 0.0385 {0.8444}

0.2580 {0.6115} 8.4484 {0.0000}

***

PHPEUR H2 H0 H1 H2 0.7787 17.2100 *** 0.1112 13.8930 * {0.8445} {0.0000} {0.1694} {0.0940} 3.1963 -8.0559 * 1.5274 -9.7304 *** {0.1677} {0.0662} {0.3132} {0.0001} -0.9450 -32.9129 * -24.5485 *** -6.1175 {0.8346} {0.0988} {0.0000} {0.5081} 0.2582 8.0148 -2.0282 * -3.3916 {0.9090} {0.3590} {0.0784} {0.2651} Tests of equality between ECBNews_R and ECBNews_F 0.3009 17.7183 *** 0.9016 7.4446 *** {0.7679} {0.0000} {0.3424} {0.0064} 0.0871 3.7088 * 3.2935 *** 0.0691 {0.7679} {0.0541} {0.0000} {0.7926} 2.2760 {0.1314} 1.2859 {0.2568}

3.0698 *** {0.0000} 4.1403 *** {0.0000}

1.8099 {0.1785} 11.1204 *** {0.0009}

Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

157

6.5.3. The Feds and the ECBs news effect on the USDEUR pair In order to be consistent with the literature, investigations of the Feds and the ECBs news effect on the mean and volatility of the USD/EUR pair are conducted. Consistent with Ehrmann and Fratzschers (2005b)s findings, Panel A of Table 6.6 shows that an unexpected hike in the Feds target interest rate leads to a significant appreciation of the USD against the EUR in all three horizons with the strongest impact occurs during the intraday horizon and the least impact occurs overnight. Furthermore, the Feds news lowers the conditional volatility of the USD/EUR pair overnight only. On the other hand, the ECBs news has no significant effect on both the conditional mean and volatility of the USD/EUR pair. The controlling for the joint effect of both the Feds and the ECBs news suggest that the separate impacts are robust (Panel B). While the ECBs news still has no significant effects, the Feds news effect on the mean is marginally weaker while the volatility impact is much stronger. Table 6.6 - The Feds and the ECBs news effects on the USD/EUR pair
This table reports the estimation results for daily changes in USDEUR exchange rate against the Feds and the ECBs news, respectively. H0 is daily horizon over the close on calendar day t-1 to the close on calendar day t; H1 is overnight horizon over the close of day t-1 to the open on day t; and H2 is intraday horizon measured over the open on day t to the close on day t. P-values are in braces. Panel A Separate effect

Conditional mean equation H0 0.9840 {0.0000} FedNews H1 0.0396 {0.0226} FedNews H1 -0.4700 {0.0000} H2 H0 0.9340 *** 1.6066 {0.0001} {0.3030} Conditional variance equation H2 0.3428 {0.6503} H0 1.9878 {0.4504} ECBNews H1 0.2767 {0.2464} ECBNews H1 -3.6999 {0.4126} H2 2.0748 {0.2942}

***

**

H0 0.0494 {0.8974}

***

H2 1.0739 {0.7432}

158

Table 6.6 Continued

Panel B Joint spillover effect Conditional mean equation H0 0.9576 {0.0000} FedNews H1 0.0328 {0.0001} H2 H0 0.9099 *** 1.0945 {0.0007} {0.6095} Conditional variance equation ECBNews H1 0.1264 {0.3896} H2 1.4775 {0.5373}

***

***

ECBNews FedNews H0 H1 H2 H0 H1 -0.0019 -2.0554 *** 0.2326 3.8038 -0.4073 {0.9968} {0.0000} {0.7030} {0.3833} {0.9017} Note: *, **, *** denotes significance at 10%, 5%, and 1%, respectively

H2 5.0298 {0.2114}

6.6. Conclusion This chapter documents and discusses the existence and the nature of the spillover effects of the policy interest rate news from the US Fed and the ECB on the first two moments of the market returns of the Asia Pacific exchange rates against the USD and the EUR for the period from January 1999 to December 2006. Considering that these are the two most important economic blocs in the world and their monetary policy news have been reported to elicit significant market movements not only in their respective markets but also in others, the comprehensive investigation of the spillover effects of these interest rate news in the Asia Pacific region adds to the literature. In general, it is revealed that the spillover effects of the Feds and the ECBs news on the Asia Pacific currencies are generally consistent with the literature, where a majority of currencies depreciates against the USD and the EUR in response to an unexpected hike in the Feds and ECBs target rates in line with the literature. More importantly, significant volatility response to the interest rate news was reported. Unexpected hikes in the Feds and the ECBs target rates reduced volatility in most of the Asia Pacific currencies against the USD and the EUR.

159

Furthermore, some evidence has been found that the news is absorbed slowly. In addition, even though the ECBs interest rate decisions may have been influenced by those of the Fed, the ECB news was also important informational provider for the Asia Pacifics currencies. The Feds news seems to send a leading signal upon the upcoming decision of the ECB, while the ECBs news tends to confirm the Feds decision. This relationship between the two news helps to reduce volatility in the Asia Pacific exchange rates against the USD and the EUR. This chapters findings provide useful information to the Asia Pacific monetary authorities in designing and implementing their own monetary policies in response to the Feds and the ECBs news. For instance, understanding that unexpected hikes in the Feds and the ECBs target interest rates would result in a depreciation of domestic currencies against the USD and the EUR, to maintain the target exchange rates, the monetary authorities of exchange rate targeting countries in the Asia Pacific region would also react accordingly by raising their target interest rates. The findings are also useful to currency traders in designing forward looking trading strategies. For example, anticipating that the Fed will increase its target interest rate higher than the markets expectation, currency traders would make profits by buying USD before the FOMCs upcoming meeting and selling back after the meeting. Furthermore, these findings also shed light on the lead-lag relationship between the Feds and the ECBs decision making and confirm our earlier findings in Chapter 5 that despite being influenced by the Feds news this study provides evidence that the ECBs news has its own merits in the Asia Pacific markets and helps resolve differences in beliefs among market participants.

160

CHAPTER 7 - CONCLUSION AND DIRECTIONS FOR FURTHER RESEARCH

161

7.1. Conclusion Over the recent past, national financial markets have become increasingly integrated and monetary policies have been important drivers for the financial linkages across markets. However, the presence and nature of this integration and the possible transmission channels are not well understood. This thesis provides comprehensive evidence on the impacts of the RBAs target news on the entire spectrum of Australian markets. Moreover, the thesis documents and discusses the spillover impacts of monetary policy news from the Fed and the ECB to the stock and the currency markets in the Asia Pacific region. Chapter 4 focused on the impacts of the RBAs target interest rate news on its domestic markets for the period from 1998-2006. This chapter contributes to the literature by providing comprehensive empirical evidence for the impacts of the RBAs target rate news on various segments of the Australian financial markets, namely debt, stock and foreign exchange markets. The RBAs news had a significant impact on the first moment of market returns/changes in line with a priori expectations, and the conditional volatility in most of the markets was significantly higher following the news. Asymmetric news effect is also observed for the Australian interest rate changes where markets tended to respond more strongly to unexpected rate rises than rate cuts. Furthermore, this study is the first research providing evidence for the Feds spillover effect on the Australian markets. We show that while the US Feds news had a first moment influence on the USD/AUD exchange rate only, the Australian market volatility, in general, was significantly lower in all market segments following the Feds news. These findings are important not only to the RBAs decision makers to improve its conducts of monetary policy

162

but also to market participants in designing trading mechanisms as well as risk management strategies in response to both domestic and external interest rate shocks. Notwithstanding the efforts to increase the transparency and effectiveness of its monetary policy, the RBAs target rate news have been found to raise volatility in most of Australian financial markets while the Feds news help to lower such volatility. This important finding has strong implications for monetary policy makers at the RBA to review and revise their policy frameworks to anchor market participants expectations on the future paths of its monetary policy stance and reduce the impacts of policy news in the markets to the extent possible. This chapters findings are also useful to market participants in designing trading mechanisms as well as risk management strategies in response to both domestic and external interest rate shocks. For example, understanding that an unexpected increase in the RBAs target interest rate would raise the 3-year Commonwealth bond interest rates, market participants would make profits if they could short-sell the 3-year bond before the RBAs meeting day and buy back at a lower price as a result of a higher interest rate. In addition, interest rate risk is perceived as the biggest risk that most of financial institutions faces due to the inherent mismatches in the assets and liabilities structure, forward looking hedging strategies building on the understanding of potential impacts of the central banks target rate would help financial institutions not only to reduce the likelihood of making loss but also increase the chance of making profits. The second investigation in Chapter 5 provided comprehensive evidence for the spillover effects of the Feds and the ECBs target interest rate news on the market returns and volatilities of twelve Asia Pacific stock markets for the period from 1999-2006. This chapter also takes an innovative approach in modeling the
163

speed of news absorption in the Asia Pacific stock markets by breaking down the close-to-close horizon into overnight and intraday horizons. This enables the investigation to capture the Asia Pacific stock markets first reaction to the news (overnight returns) and any delayed reactions during the trading day in the Asia Pacific region (intraday returns). We show that both the target rate news have negative impacts on the returns and raise volatility in most of the Asia Pacific stock markets, and the news is absorbed slowly. These findings have important implications for the Asia Pacifics policy makers and market participants alike in anticipating and managing potential spillover effects from the Feds and the ECBs so as to have forward looking policy actions and hedging strategies. For instance, the notion that the ECB follows the Fed in setting its policy is so strong amongst market participants, there might be argument that policy makers and market participants need not to consider the ECBs target rate actions but keep a close watch at the Fed only. We, however, provided evidence that the ECB news was also another important informational leader that Asia Pacific region should follow closely as the ECBs impacts on the Asia Pacific markets are consistent with those imposed by the Feds decisions. Furthermore, understanding that unexpected hikes in the Feds and the ECBs target rates would result in a negative return in most of the Asia Pacific stock indices, index traders would also design their trading and hedging strategies accordingly to shield the risk and make profits if their expectations are correct. The third study in Chapter 6 examined the spillover impacts of the Feds and the ECBs target interest rate news on the first two moments of seven Asia Pacific exchange rates against the US dollar and the euro over the period from 1999-2006. Consistent with the literature, a majority of Asia Pacific currencies depreciates against the USD and the EUR in response to the Feds and the ECBs unexpected
164

rate rises. Both the Fed and the ECB news elicited delayed or persisting volatility responses. More importantly, the ECBs news tends to confirm the Feds decisions. This relationship between the news tends to help reduce volatility in the Asia Pacific currency markets. This chapters findings provide useful information to the Asia Pacific monetary authorities in designing and implementing their own monetary policies in response to the Feds and the ECBs news. For instance, understanding that unexpected hikes in the Feds and the ECBs target interest rates would result in a depreciation of domestic currencies against the USD and the EUR, to maintain the target exchange rates, the monetary authorities of exchange rate targeting countries in the Asia Pacific region would also react accordingly by raising their target interest rates. The findings are also useful to currency traders in designing forward looking trading strategies. For example, anticipating that the Fed will increase its target interest rate higher than the markets expectation, currency traders would make profits by buying USD before the FOMCs upcoming meeting and selling back after the meeting. Furthermore, these findings also shed lights on the lead-lag relationship between the Feds and the ECBs decision markings and confirm our earlier findings in Chapter 5 that despite being influenced by the Feds news this study provides evidence that the ECBs news has its own merits in the Asia Pacific markets and helps resolve differences in beliefs among market participants. In short, this thesis provides better understandings on not only the potential impacts of the RBAs target interest rate news on both the conditional mean and volatility of Australian financial market returns; but also shed lights on the spillover effects of target interest rates news from the two most important central banks, the Fed and the ECB, on the Asia Pacific financial markets.

165

7.2. Directions for future research Despite comprehensive investigation results provided in this thesis, there is a scope for further research. This section discusses some of the possible extensions. 7.2.1. Higher frequency data The first empirical investigation on the domestic effects of the RBAs news on Australian financial markets uses daily data as Faust et al. (2003) identify that asset prices may initially demonstrate overshooting reaction to news. Thus, the permanent effect of news may be smaller than the initial reaction and the news may be incorporated gradually over several minutes or even hours after the actual announcement. However, high frequency investigation of the spillover impacts of the Feds and the ECBs target rate news is not possible as the Feds and the ECBs announcements are made when the Asia Pacific markets are closed. As such, the earliest market reaction on these news events can be captured at market opening one calendar day after the news announcement. The development of high frequency data allows for empirical investigations of a wide range of issues in the financial markets. Therefore, the use of higher frequency data, i.e. intraday data, to investigate the impacts of target rate news on domestic markets would provide more accurate evidence for the transmission of monetary policy effects into the economy where the problem of omitted variables is partially addressed. 7.2.2. Sectoral and firm level effects In Chapter 5 and 6 we show comprehensive evidence for the Feds and the ECBs news effects on the Asia Pacific stock markets at aggregate levels, i.e. stock

166

indices. One would expect that different industry sectors within the aggregate indices would react differently to the target rate news. There are some studies on the impacts of the target rate news on stock markets at sectoral levels, including the banking sector, for example Bredin et al.(2009) investigate the ECBs news effects on the UKs and Germans aggregate and sectoral stock returns while Basistha and Kurov (2008) investigate the Feds news effects at the US sub-indices. An investigation for the spillover impacts of the Feds and the ECBs target rate news on different sectors of the Asia Pacific stock markets would significantly contribute to the literature. In addition, we show in Chapter 4 that Australian banks respond significantly to the target rate news. As the Asia Pacific banks are perceived to be one of the most important conduits of monetary policy and they are main providers for fund to most of firms in the region, an in-depth analysis for the Feds and the ECBs spillover effects on the first two moments of the Asia Pacific bank stocks at the firm level would also make significant contributions to the literature. 7.2.3. Determinants of the spillover effects We show in Chapters 5 and 6 that markets in the Asia Pacific region respond differently to the news from the Fed and the ECB, however, we do not empirically investigate the factors behind these differences. The Asia Pacific region is diverse in terms of the degrees of openness and institutional arrangements, e.g. monetary and exchange rate regimes. Furthermore, different markets would also have different degrees of financial and real integration with the US and the euro area in terms of trade and investments. Such diversity might play a role in explaining for the crossmarket differences. Therefore, a deeper investigation of domestic macroeconomic policies as well as the degrees of financial and real integration between the Asia

167

Pacific markets and the US and euro area markets would provide a better understanding on the determinants of spillover effects. 7.2.4. Other macroeconomic news spillover impacts The controlling for other macroeconomic announcements from the U.S. and the euro area in the investigations in Chapters 5 and 6 seems to reveal that these important releases also have significant impacts on the volatility of the Asia Pacific markets. Furthermore, Kim et al (2004) find that the USs macroeconomic news have significant impacts on the US bond, stock and foreign exchange markets. Therefore, an extension to investigate the role of the USs and the euro areas macroeconomic news on the Asia Pacific markets would make further contributions to the literature by providing evidence for (i) which macroeconomic news from the US and euro area have significant influences on the Asia Pacific markets; (ii) the direction and magnitude of such impacts; and (iii) among the US and euro area, who would lead the Asia Pacific region. As the world is becoming smaller due to globalization and integration, geographical distance might no longer be an issue and financial and real integration would play a key role in explaining the spillover impacts. This thesis and some possible extensions outlined above are just the beginning of the prolonged crusades in search for the Holy Grail of the nature and determinants for the spillover effects of the target interest rate news from a central bank to foreign financial markets.

168

APPENDICES

169

Appendix A

Australian macroeconomic announcements made on the same date as the RBA's target rate announcements
Macroeconomic announcement CPI Inflation (Cat. No. 6403) Announcement date 08-Aug-01 07-Nov-01 06-Feb-02 08-May-02 03-Jul-02 07-Aug-02 06-Nov-02 07-May-03 05-Nov-03 02-Sep-04 02-Feb-05 04-May-05 03-Aug-05 02-Nov-05 03-May-06 02-Aug-06 07-Apr-99 05-Apr-00 05-Jul-00 04-Oct-00 02-Oct-02 03-May-00 03-Apr-02 08-Jun-05 08-Feb-06 08-Mar-06 04-Oct-06 08-Nov-06 04-Feb-98 06-May-98 04-Aug-99 03-Nov-99 02-Aug-00 03-Oct-01 02-Apr-03 02-Jul-03 06-Aug-03 03-Sep-03 04-Feb-04 03-Nov-04 04-May-05 02-Aug-06

Employment/Unemployment (Cat. No. 6202.2)

International Accounts (Cat. No. 5368)

Retail sales (Cat no. 8501)

Note: The Australian Bureau of Statistics publications (i.e. announcements) are released to the general public at 11:30 am Australian EST on the day of announcement and are embargoed until the official release.

170

Appendix B

Granger causality tests for the Fed and the RBA

This table reports the Granger causality tests results for the hypotheses that the Fed Granger causes the RBA in target rates decision making process (the left panel), and the RBA Granger causes the Fed in target rates decision making process (the right panel). P-values are in braces. Fed Granger causes RBA RBA Granger causes Fed Lags F-value Lags F-value 1 4.7639 ** 1 0.00001 {0.0292} {0.9982} 2 2.3790 * 2 0.00001 {0.0929} {1.0000} 3 1.5840 3 0.00001 {0.1912} {1.0000} 4 1.1880 4 0.2964 {0.3140} {0.8805} 5 4.7718 *** 5 0.4743 {0.0002} {0.7957} Note: *, **, *** denote significance at 10%, 5%, and 1%, respectively

171

Appendix C

The US market responses to the US Feds target interest rate news Chapter 4 has demonstrated the existence of a direct influence of the US Feds target interest rate announcement news on the Australian financial markets. In this Appendix, a comprehensive empirical evidence of the extent to which the Feds news has an impact on the US financial markets is reported. The coverage of the US markets includes debt, foreign exchange and stock markets. Panel A of Table C.1 shows the daily US market returns data which consist of short- and long-term interest rates (90-day Treasury bills, 3- and 10-year Treasury bonds); spot and 1- and 3-month forward exchange rates of the US dollar against the Euro (EUR), Japanese yen (JPY) and British pound (GBP); and the two largest stock indices, the S&P Composite and the Dow Jones Industrial Average (DJIA). These are all sourced from Datastream. In most cases, significant negative skewness and serial correlation in returns were observed. In addition, all cases show leptokurtosis, non-normality and significant serial correlation in the second moments. Panel B of Table C.1 reports the estimation results of the Feds news effects on the US financial markets. For the debt market, the short-term interest rate

responded stronger to the news than the longer-term rates where the 90-day rate increased by 0.0562 percentage point in response to a one percentage point unexpected rise in the Feds target rate, while the 10-year rate fell by 0.0125 percent. For the foreign exchange market, only the 1-month forward rate responded. The USD/JPY 1-month forward rate depreciated by 0.0838 percent and the USD/EUR appreciated by 0.0272 percent in response to a one percentage point unexpected rise in the Feds target rate.
172

The news has reduced the conditional volatility, in general. The conditional volatilities of the USD/EUR, the USD/GBP exchange rate returns and stock index returns are significantly lower in response to the news. However, the target interest rate news significantly increased the conditional volatility of the daily changes of the 90-day Treasury bill rate. This suggests that the new information contained in the Feds target interest rate announcements helped to reduce the level of heterogeneity in belief among market participants regarding the future course of the US economy.

173

Table C.1 - The US financial market returns and the US Feds target interest rate news effects

This table reports descriptive statistics for the US financial market returns/changes series (Panel A), and the estimation results for the impacts of the USs Fed news on the US financial markets (Panel B)

Panel A reports the summary statistics of daily changes of short-and long-term interest rates, daily returns of spot and 1-month and 3-month forward USD exchange rates against the Euro, Yen and Pound Sterling, and daily returns of two stock market indices (S&P composite and Dow Jones Industrial Average)

Panel B reports the news coefficients in the conditional mean and variance of the EGARCH(1,1) equations employed to examine the Feds news effects on the US financial markets as shown below. The dependant variable, yt, is the daily US financial market variables shown in Panel A. P-values are in braces.
p

y t = c + Lag ,i y t i + Mon Mont + Hol Hol t + FedNews FedNewst + t

i =1

ln ht = c + h ln ht 1 + 1 ht 1 ht 1

t 1
+ 2 + Mon Mon t + Hol Hol t + FedNews FedNews t

t 1

Panel A. The US financial market returns

Stock Market 3 m forward 2370 -0.0032 0.7335 -0.7399 7.6520 5996
*

Debt markets Spot 2087 -0.0056 0.6017 -0.2110 0.9802 99 USDEUR 1 m forward 2087 -0.0056 0.6005 -0.2005 0.9568 94 3 m forward 2087 -0.0055 0.5998 -0.1965 0.9449 91 Spot 2370 -0.0034 0.6880 -0.7392 6.0666 3849

90-day T-bill 3-Year T-bond 10-year T-bond 2370 2370 2370 -0.0002 -0.0004 -0.0005 0.0443 0.0607 0.0512 -1.7206 0.0959 0.2889 28.2719 3.0337 1.1118 80066 912 155

Exchange rates USDJPY 1 m forward 2370 -0.003263 0.732554 -0.733938 9.884944 9858

Spot 2370 -0.0063 0.4925 -0.1080 0.7649 62

*

USDGBP 1 m forward 2370 -0.0064 0.4921 -0.1016 0.7521 60

**

3 m forward SP Composite Dow Jones 2370 2370 2370 -0.0065 0.0158 0.0186 0.5372 1.1258 1.0875 -0.2723 -0.0016 -0.1252 6.4683 2.9089 3.9307 4159 835 1531
**

***

***

No. of observations Mean Std. deviation Skewness Kurtosis Jarque-Bera Ljung-Box Q test Return P-Value Volatility P-Value 23.8262 {0.2501} 161.7333 {0.0000}
***

120.6239 {0.0000} 181.2865 {0.0000} 10.2505 {0.9634} 51.3407 *** {0.0001} 10.0663 {0.9670} 53.6618 *** {0.0001} 10.0669 {0.9669} 55.0158 *** {0.0000}

***

39.0931 {0.0065} 162.3160 {0.0000}

***

28.5205 {0.0976} 464.2989 {0.0000}

***

24.9247 0.204327096 183.7811 *** {0.0000}

33.8427 ** {0.0272} 199.2567 *** {0.0000}

31.1707 {0.0530} 117.4211 {0.0000}

***

33.7111 {0.0281} 118.5834 {0.0000}

***

31.0644 * {0.0543} 303.9345 *** {0.0000}

32.6041 {0.0373} 989.3466 {0.0000}

***

26.3564 {0.1544} 839.8300 {0.0000}

***

174

Table C.1 - Continued

Panel B. The effects of the Fed's target rate news

Stock Market

Debt markets

FedSur

***

FedSur

90-day 0.0562 {0.0000} 0.4505 {0.0000} Spot 0.0710 {0.2606} -0.0469 {0.2880} Spot 0.1076 {0.1149} -0.7318 {0.0000}

3 Year 0.0042 {0.7449} *** -0.0255 {0.7864}

10 year -0.0125 {0.0661} 0.1246 {0.1288}

Spot -0.0245 {0.1458} -1.4407 {0.0000}

USDEUR 1 m forward3 m forward -0.0272 * -0.0251 {0.0627} {0.1198} *** -1.4436 *** -1.4294 *** {0.0000} {0.0000}

Exchange rates USDJPY 1 m forward 3 m forward 0.0838 ** 0.0812 {0.0475} {0.1185} -0.0827 -0.0332 {0.2023} {0.3092} USDGBP 1 m forward 3 m forward 0.1079 -0.0756 {0.1096} {0.5406} *** -0.7488 *** 0.1594 {0.0000} {0.7855}

SP Composite Dow Jones -0.1540 -0.1157 {0.1771} {0.3376} -0.2126 *** -0.2808 *** {0.0010} {0.0006}

Note: *, **, *** denote significance at 10%, 5%, and 1%, respectively

175

Appendix D Controlled macroeconomic announcements

U.S' macroeconomic announcements Euro area's macroeconomic announcements

Announcement types Total Total

No. of announcements on the same dates as the Fed's target rate announcements 1 2 0 8 2 Jan 2002 - Dec 2006 66 112 59

No. of announcements on the same dates as the ECB's target rate announcements 12 2 6

Jan 1999 - Dec 2006 32 96 96 96 95

Sample period Gross Domestic Product (GDP) Consumer Price Index (CPI) Unemployment (EU) Trade deficit (TD) Retail sales (RET) External trade (ET) Current account balance (CAB) Retail trade growth rate (RTT) Total Proportion to total announcements Proportion to total Fed's/ ECB's announcements 415 13 3.13% 18.84% 62 36 59 394

0 0 7 27 6.85% 20.61%

176

REFERENCES
Andersen, T.G., Bollerslev, T., Diebold, F.X., Vega, C., 2003. Micro effects of macro announcements: real-time price discovery in foreign exchange. American Economic Review 93, 38-62. Andersen, T.G., Bollerslev, T., Diebold, F.X., Vega, C., 2007. Real-time price discovery in stock, bond and foreign exchange markets. Journal of International Economics 73, 251-277. Arshanapalli, B., Doukas, J., Lang, L.H.P., 1995. Pre and post-October 1987 stock market linkages between US and Asian markets. Pacific Basin Finance Journal 3, 57-73. Basistha, A., Kurov, A., 2008. Macroeconomic cycles and the stock markets reaction to monetary policy. Journal of Banking and Finance 32, 2606-2616. Bernanke, B.S., 2003. A Perspective on Inflation Targeting. Remarks at the Annual Washington Policy Conference of the National Association of Business Economists, Washington DC, 25 March. Bernanke, B.S., Blinder, A.S., 1992. The Federal Funds Rate and the Channels of Monetary Transmission. American Economic Review 82, 901-921. Bernanke, B.S., Gertler, M., 1995. Inside the Black Box: The Credit Channel of Monetary Policy Transmission. Journal of Economic Perspectives 9, 27-48. Bernanke, B.S., Gertler, M., Gilchrist, S., 1996. The financial accelerator and the flight to quality. Review of Economics and Statistics 78, 1-15. Bernanke, B.S., Kuttner, K.N., 2005. What explains the stock markets reaction to Federal Reserve policy? Journal of Finance 60, 1221-1257.
177

Bernoth, K., Von Hagen, J., 2004. Euribor futures market: efficiency and the impact of ECB policy announcements. International Finance 7, 1-24. Bindseil, U., 2002. Equilibrium bidding in the Eurosystem's open market operations. European Central Bank Working paper series No. 137. Black, F., 1976. Studies of stock price volatility changes. Proceedings of the 1976 Meetings of the American Statistical Association, Business and Economics Statistics Section, 177-181. Blanchard, O., Rhee, C., Summers, L., 1993. The stock market, profit, and investment. Quarterly Journal of Economics 108,115136. Blinder, A.S., Ehrmann, M., Fratzcher, M., de Hann, J., Jansen, D.J., 2008. Central bank communication and monetary policy: A survey of theory and evidence. Journal of Economic Literature 46, 910-945. Bjornland, H.C., Leitemo, K., 2009. Identifying the interdependence between US monetary policy and the stock market. Journal of Monetary Economics 56, 275-282. Bohl, M.T., Siklos, P.L., Sondermann, D., 2008. European Stock Markets and the ECBs Monetary policy Surprises. International Finance 11, 117-130. Bomfim, A.N., 2003. Pre-announcement effects, news effects, and volatility: Monetary policy and the stock market. Journal of Banking and Finance 27, 133-151. Bond, S., Cummins, J., 2001. Noisy share prices and the Q Model of investment. Institute for Fiscal Studies Working papers No. W01/22.

178

Bredin, D., Gavin, O., O'Reilly, G., 2005. US monetary policy announcements and Irish stock market volatility. Applied Financial Economics 15, 1243 1250. Bredin, D., Hyde, S., Nitzsche, D., O'Reilly, G., 2009. European monetary policy surprises: the aggregate and sectoral stock market response. International Journal of Finance & Economics 14, 156-171. Brunner, A.D., 2000. On the derivation of monetary policy shocks: Should we throw the VAR out with the bath water? Journal of Money, Credit and Banking 32, 254279. Christie, A.A., 1982. The Stochastic Behavior of Common Stock Variances: Value, Leverage and Interest rate effects. Journal of Financial Economics 10, 407432. Concover, C.M., Jensen, R.G., Johnson, R., 1999. Monetary environments and international stock returns. Journal of Banking and Finance 23, 13571381. Connolly, R.A., Wang, F.A., 2003. International equity market comovements: Economic fundamentals or contagion? Pacific Basin Finance Journal 11, 2343. Cook, T., Hahn, T., 1989. The effects of changes in the Federal funds rate target on market interest rates in the 1970s. Journal of Monetary Economics 24, 331351. Demiralp, S., Jorda, O., 2004. The response of term rates to Fed announcements. Journal of Money, Credit and Banking 36, 387-405. Diggle, J., Brooks, R., 2007. The target cash rate and its impacts on investment asset returns in Australia. Applied Financial Economics 17, 615-633.

179

Driscoll, J., 2004. Does bank lending affect output? Journal of Monetary Economics 51, 451471. Durham, J.B., 2001. The effect of monetary policy changes on monthly and quarterly stock market returns: cross country evidence and sensitivity analysis. Board of Governors of the Federal Reserves System Finance and Economics Discussion Series 2001-42. Edey, M., Elliott, G., 1988. Pricing Behaviour in Australian Financial Future Markets. Reserve Bank of Australia Research discussion papers No. rdp8804. Ehrmann, M., Fratzscher, M., 2003. Monetary policy announcements and money markets: a transatlantic perspective. International Finance 6, 309-328. Ehrmann, M., Fratzscher, M., 2004. Taking stock: Monetary policy transmission to equity markets. Journal of Money, Credit, and Banking 36, 719737. Ehrmann, M., Fratzscher, M., 2005a. Equal size, equal role? Interest rate interdependence between the euro area and the United States. Economic Journal 115, 928948. Ehrmann, M., Fratzscher, M., 2005b. Exchange rates and fundamentals: new evidence from real-time data. Journal of International Money and Finance 24, 317-341. Ehrmann, M., Fratzscher, M., 2006. Global financial transmission of monetary policy shocks. European Central Bank Working paper series No. 616. Eichenbaum, M., Evans, C.L., 1995. Some empirical evidence on the effects of shocks to monetary policy on exchange rates. Quarterly Journal of Economics 110, 9751009.

180

Engle, R.F., Ng, V., 1993. Measuring and Testing the impact of News on Volatility. Journal of Finance 48, 1749-1778. Faff, R.W., Hillier, D., McKenzie, M.D., 2009. Modelling return and volatility exposures in the global stock markets. Global Finance Journal, Forthcoming. Fatum, R., Scholnick, B., 2008. Monetary policy news and exchange rate responses: Do only surprises matter? Journal of Banking and Finance 32, 1076-1086. Faust, J., Rogers, J.H., 2003. Monetary policy's role in exchange rate behavior. Journal of Monetary Economics 50, 1403-1424. Faust, J., Rogers, J.H., Swanson, E., Wright, J.H., 2003. Identifying the effects of monetary policy shocks on exchange rates using high-frequency data. Journal of the European Economic Association 1, 10311057. Fratzscher, M., 2008. US shocks and global exchange rate configurations. Economic Policy 23, 363-409. French, K., Schwert, G.W., Stambaugh, R.F., 1987. Expected stock returns and volatility. Journal of Financial Economics 19, 3-29. Gasbarro, D., Monroe, G.S., 2004. The impact of monetary policy candidness on Australian financial markets. Journal of Multinational Financial Management 14, 35-46. Gaspar, V., Perez-Quiros, G., Sicilia, J., 2001. The ECB monetary policy strategy and the money market. International Journal of Finance & Economics 6, 325342.

181

Ghosh, A., Saidi, R., Johnson, K.H., 1999. Who moves the Asia Pacific stock markets US or Japan? Empirical evidence based on theory of cointegration. The Financial Review 34, 159-170. Gilchrist, S., Leahy, J.V., 2002. Monetary policy and asset prices. Journal of Monetary Economics 49, 75-97. Glosten, L.R., Jagannathan, R., Runkle, D.E., 1993. On the relation between the expected value and volatility of the nominal excess returns on stocks. Journal of Finance 48, 1779-1801. Gulley, O.D., Sultan, J., 2003. The link between monetary policy and stock and bond markets: evidence from the federal funds futures contract. Applied Financial Economics 13, 199-209. Grkaynak, R.S., Sack, B., Swanson, E., 2007. Market-Based Measures of Monetary Policy Expectations. Journal of Business and Economic Statistics 25, 201212. Haldane, A., Read, M., 2000. Monetary Policy Surprises and the Yield Curve. Bank of England Working Paper No. 106. Hamao, Y., Masulis, R., Ng, V., 1990. Correlations in price changes and volatility across international stock markets. Review of Financial Studies 3, 281-307. Hausman, J., Wongswan, J., 2006. Global asset prices and FOMC announcements. Board of Governors of the Federal Reserve System International Finance Discussion Papers No. 886. Henry, O.T., Summers, P.M., 1999. The Volatility of Real Exchange Rates: The Australian Case. Australian Economic Papers 38, 79-90.

182

Hogan, L.I., Nguyen, D.T., Exchange rate target zones and the conduct of monetary policy: an Australian perspective. Australian National University, Centre for Economic Policy Research Discussion paper No. 184, 33-37. Joiron, P., 1996. Risk and turnover in the foreign exchange market. In: Frankel, J., Galli, G., Giovannini, A. (Eds.), The Microstructure of Foreign Exchange Markets. The University of Chicago Press, Chicago, 1940. Juttner, D.J., Tuckwell, R.H., Luedecke, B.P., 1985. Are Expectations of Short-term Interest Rates Rational? Australian Economic Papers 24, 356-369. Kashyap, A.K., Stein, J.C., 2000. What do a million observations on banks say about the transmission of monetary policy? American Economic Review 90, 407428. Kashyap, A.K., Stein, J.C., Wilcox, D.W., 1993. Monetary policy and credit conditions: Evidence from the composition of external finance. American Economic Review 83, 78-98. Kearney, C., 2000. The determination and international transmission of stock market volatility. Global Finance Journal 11, 31-52. Kim, S.-J., 2003. The spillover effects of US and Japanese public information news in advanced Asia Pacific stock markets. Pacific Basin Finance Journal 11, 611-630. Kim, S.-J., 2005. Information leadership in the advanced Asia Pacific stock markets: Return, volatility and volume information spillovers from the US and Japan. Journal of the Japanese and International Economies 19, 338-365.

183

Kim,

S.-J.,

McKenzie,

M.D.,

Faff,

R.W.,

2004.

Macroeconomic

news

announcements and the role of expectations: evidence for US bond, stock and foreign exchange markets. Journal of Multinational Financial Management 14, 217-232. Kim, S.-J., Nguyen, D.Q.T., 2008. Target Interest Rate News Spillover Effects on the Asia-Pacific Currency Markets. International Research Journal of Finance and Economics 20, 27-45. Kim, S.-J., Nguyen, D.Q.T., 2008. The Reaction of the Australian Financial Markets to the Interest Rate News from the Reserve Bank of Australia and the US Fed. Research in International Business and Finance 22 (3), 378-395. Kim, S.-J., Nguyen, D.Q.T., 2009. The Spillover Effects of Target Interest Rate News from the U.S. Fed and the European Central Bank on the Asia-Pacific Stock Markets. Journal of International Financial Markets, Institutions & Money 19 (3), 415-431. Kim, S.-J., Sheen, J., 2000. International linkages and macroeconomic news effects on interest rate volatility - Australia and the US. Pacific Basin Finance Journal 8, 85-113. Kim, S., Roubini, N., 2000. Exchange rate anomalies in the industrial countries: a solution with a structural VAR approach. Journal of Monetary Economics 45, 561586. Krueger, J., Kuttner, K., 1996. The Fed funds futures rate as a predictor of Federal Reserve policy. Journal of Futures Markets 16, 865-879. Kuttner, K.N., 2001. Monetary policy surprises and interest rates: Evidence from Fed funds futures market. Journal of Monetary Economics 47, 523-544.
184

Lee, J., 2006. The impact of federal funds target changes on interest rate volatility. International Review of Economics and Finance 15, 241-259. Lewis, K.K., 1995. Are foreign exchange intervention and monetary policy related and does it really matter? Journal of Business 68, 185-214. Liu, Y.A., Pan, M.S., 1997. Mean and Volatility spillover effects in the US and Pacific Basin stock markets. Multinational Finance Journal 1, 747-762. Liu, Y.A., Pan, M.S., Shieh, J.C.P., 1998. International transmission of stock price volatility: Evidence from the US and five Pacific Basin markets. Journal of Economics and Finance 22, 56-69. Lobo, B.J., Darrat, A.F., Ramchander, S., 2006. The Asymmetric Impact of Monetary Policy on Currency Markets. The Financial Review 41, 289-303. Masih, A.M.M., Winduss, T., 2006. Who leads the Australian interest rates in the short and long run? An application of long run structural modeling. Review of Pacific Basin Financial Markets and Policies 9, 1-24. Miron, J., Romer, C.D., Weil, D., 1994. Historical perspectives on the monetary transmission mechanism. In: Mankiw NG (ed.) Monetary Policy. University of Chicago Press, Chicago, IL. Mishkin, F.S., 1996. The Channels of Monetary Transmission: Lessons for Monetary Policy. Banque de France Bulletin Digest 27, 33-44. Miyakoshi, T., 2003. Spillovers of stock return volatility to Asian equity markets from Japan and the US. Journal of International Financial Markets, Institutions and Money 13, 383-399.

185

Modigliani, F., 1971. Monetary Policy and Consumption. In: Consumer Spending and Monetary Policy: The Linkages. Federal Reserve Bank of Boston Conference Series 5, 9-84. Narayan, P.K., Smyth, R., 2004. Modeling the linkages between the Australian and G7 stock markets: Common stochastic trends and regime shifts. Applied Financial Economics 14, 991-1004. Nelson, D.B., 1991. Conditional Heteroscedasticity in Asset Returns: A new approach. Econometrica 62, 1-41. Newman, G., 2006. Rates to keep the markets nervous. The Australian, 6 November 2006, 33. Ng, A., 2000. Volatility spillover effects from Japan and the US to the Pacific Basin. Journal of International Money and Finance 19, 207-233. Peersman, G., Smets, F., 2005. The industry effects of monetary policy in the Euro area. Economic Journal 115, 319-342. Perez-Quiros, G., Sicilia, J., 2002. Is the European Central Bank (and the United States Federal Reserve) predictable? European Central Bank Working paper series No. 192. Phylaktis, K., Ravazzolo, F., 2005. Stock market linkages in emerging markets: implications for international portfolio diversification. Journal of

International Financial Markets, Institutions and Money 15, 91-106. Reinhart, V.R., Simin, T., 1997. The market reaction to Federal Reserve policy action from 1989 to 1992. Journal of Economics and Business 49, 149-168.

186

Rigobon, R., Sack, B.P., 2004. The Impact of Monetary Policy on Asset Prices. Journal of Monetary Economics 51, 1553-1575. Roley, V.V., Sellon, G.H., 1995. Monetary policy actions and long-term interest rates. Economic Quarterly 80, 73-89. Roley, V.V., Sellon, G.H., 1998. Market reaction to monetary policy

nonannouncement. Federal Reserve Bank of Kansas City working paper No. 98-06. Rudebusch, G., 1998. Do measures of monetary policy in a VAR make sense? International Economic Review 39, 907-931. Tauchen, G., Pitts, M., 1983. The price variabilityvolume relationship on speculative markets. Econometrica 51, 485505. Taylor, J.B., 1995. The Monetary Transmission Mechanism: An Empirical Framework. Journal of Economic Perspectives 9, 11-26. Tease, W., 1988. The Expectations Theory of the Term Structure and Short-Term Interest Rates in Australia. Economic Record 64, 120 127. Thornton, D.L., 1998. Tests of the markets reaction to Federal Funds rate target changes. Federal Reserve Bank of St. Louis Review 80, 25-36. Tobin, J., 1969. A general equilibrium, approach to monetary theory. Journal of Money, Credit, and Banking 1, 15-29. Valente, G., 2008. International interest rates and US monetary policy announcements: Evidence from Hong Kong and Singapore. Journal of International Money and Finance 28, 920-940. Vickers, J., 2000. Monetary policy and asset prices. The Manchester School 68, 1-22.
187

Warner, E.J., Georges, C., 2001. The credit channel of monetary policy transmission: Evidence from stock returns. Economic Inquiry 39, 74-85. Wongswan, J., 2006. Transmission of Information across International Equity Markets. Review of Financial Studies 9, 1157-1189. Wongswan, J., 2009. The response of global equity indices to US monetary policy announcements. Journal of International Money and Finance 28, 344-365. Wrtz, F.R., 2003. A comprehensive model on the Euro overnight rate. European Central Bank Working paper series No. 207. Wynne, M.A., 1999. The European System of Central Banks. Federal Reserve Bank of Dallas Economic and Financial Policy Review QI, 2-14.

188