بحران های مالی و ارتباط پویا میان ارزهای بین المللی
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|14827||2013||14 صفحه PDF||سفارش دهید||6930 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Journal of International Financial Markets, Institutions and Money, Volume 26, October 2013, Pages 319–332
This paper investigates the interdependence of US dollar exchange rates expressed in other major currencies. Focusing on different phases of the Global financial crisis (GFC) and the Eurozone Sovereign Debt Crisis (ESDC), we adopt a dynamic conditional correlation model into a multivariate Fractionally Integrated Asymmetric Power ARCH (FIAPARCH) framework, during the period 2004–2011. The findings indicate a decrease of exchange rates correlations during the turmoil periods, suggesting the different vulnerability of the currencies. The most stable periods of the two crises for all currencies are the early phases of the GFC, while the first phase of ESDC exhibit the most cases of decreasing correlations. Finally, the Japanese yen and Swiss franc show evidence of safe heaven currencies across several phases of the two crises. The results provide crucial implications for portfolio diversification strategies and highlight the need for some form of policy coordination among central banks.
The global severe deterioration in various asset markets as well as in macroeconomic fundamentals occurred during the last half decade, characterized it as one of the most unanticipated and tumultuous periods in the recent economic history. Two major crises events took place over this turmoil period: the Global financial crisis (GFC, hereafter) of 2007–2009 and the Eurozone Sovereign Debt Crisis (ESDC, hereafter) that started at the fall of 2009 and triggered by the Greek debt problem. During crises, the issues of risk management and asset allocation are very important to practitioners and academics. While the effects of these two crises on stock, bond and other asset markets have been described and analyzed (e.g., Arghyrou and Kontonikas, 2012, Baur, 2012, Chan et al., 2011, Guo et al., 2011, Kenourgios and Padhi, 2012 and Dimitriou et al., 2013), studies that examine the behavior of exchange rates during those turmoil periods are still rare. There is prior literature investigating the volatility spillover linkages among exchange rates since the seminal paper of Engle et al. (1990), who support that the uncertainty in exchange rates arises not only from local shocks, but also transmitted across markets. Most of the early literature analyzes the volatility spillovers among currencies employing conventional methodologies, such as cointegration, causality, generalized autoregressive conditional heteroskedasticity (GARCH) specifications and cross-correlation function. For example, Nikkinen et al. (2006) examine the expected future volatilities among major European currencies during the period 2001–2003 and find that implied volatility of EUR affects those of GBP and CHF. Inagaki (2007) examines volatility spillover effects among EUR and GBP during the period 1999–2004 and finds that EUR unidirectional causes-in-variance the GBP. However, this literature suffers from certain limitations. First, cointegration methodology does not accommodate the possibility of non-normality and asymmetry in the variance of returns (Baele, 2005). Second, there is a heteroskedasticity problem when measuring correlations caused by volatility increases during a crisis (Forbes and Rigobon, 2002). Third, most of the GARCH family models assume that correlation coefficients are constant over the sample period, while their multivariate variants suffer from the curse of dimensionality. Fourth, empirical analysis must examine the second moments of correlations and covariances in order to provide evidence of dynamic changes in linkages among markets across stable and crisis periods (Pesaran and Pick, 2007). To avoid the limitations of this literature, recent research on exchange rates linkages focuses on their dynamic conditional correlations in a time-varying GARCH framework. The dynamic conditional correlation (DCC) GARCH model developed by Engle (2002) provides a robust analysis of time-varying linkages by allowing conditional asymmetries in both volatilities and correlations, while investigates the second order moments dynamics of financial time-series and overcomes the heteroskedasticity problem.1 For example, Perez-Rodriguez (2006) provides evidence on the fluctuation of correlations among EUR, GBP and CHF over the period 1999–2004, using the DCC-GARCH model of Engle (2002). Kitamura (2010) analyzes intraday volatility spillovers among EUR, CHF and GBP via a Varying-Correlation (VC) multivariate GARCH model and finds evidence of spillover effects from EUR to the other two currencies during the period 2008–2009. Antonakakis (2012) uses a DCC-GARCH model and finds bidirectional cross-market volatility spillovers among major European exchange rates markets, focusing on the period before and after the introduction of euro.2 This work focuses on the impact of the GFC and ESDC on major international currencies. Specifically, we empirically investigate the time-varying linkages of five daily US dollar exchange rates, namely euro (EUR), Japanese yen (JPY), British pound (GBP), Swiss franc (CHF) and Australian dollar (AUD) from 1st January 2004 to 31st December 2011. While an increasing number of studies use the standard GARCH-DCC model of Engle (2002), we differentiate our analysis by employing a DCC model into a multivariate fractionally integrated APARCH framework (FIAPARCH-DCC model). In this set up, we estimate the time-varying dynamic conditional correlations among the currencies and then examine the dynamic patterns of correlation changes across four phases of the GFC and two phases of ESDC. The advantage of the FIAPARCH model of Tse (1998) used in this paper is its flexibility, since it includes a large number of alternative GARCH specifications. Specifically, it increases the flexibility of the conditional variance specification by allowing long-range volatility dependence and an asymmetric response of volatility to positive and negative shocks. Moreover, it allows the data to determine the power of returns for which the predictable structure in the volatility pattern is the strongest (Conrad et al., 2011). Finally, Conrad et al. (2011) and Chkili et al. (2012) support the forecasting superiority of FIAPARCH on other GARCH family models. From an economic perspective, monetary authorities may intervene in order to sustain price stability and competitiveness in exports, when other currencies depreciate or appreciate. This behavior may cause different degrees of exchange rates co-movements during turmoil periods compared to stable periods. Therefore, it would be useful to empirically investigate the dynamic dependence structure of international currencies during a turbulent period which covers two of the most severe crises occurred the last decades. To identify the turmoil periods and their phases, we use both official data sources for important events and regimes of excess volatility. The present study contributes to the existent literature in the following aspects. Firstly, we are among the first to examine separately time-varying correlation dynamics among currencies during different phases of the GFC and ESDC, following a crisis identification based on both financial/economic events and a statistical approach. Secondly, we provide a robust analysis of time-varying linkages among currencies that goes beyond a simple analysis of correlation breakdowns. The present work analyses the second moment dynamics of exchange rates taking into account long memory behavior, leverage effects and asymmetries. Finally, many interesting aspects emerged from this paper. The decreasing pattern of exchange rates linkages during the turmoil periods suggests the different vulnerability of the currencies. At the beginning of the GFC, the British pound shows the greater exposure, while the JPY and CHF provide hedging benefits and could be characterized as safe heaven currencies across several periods of the two crises. Across all phases of the two crises, the most cases of decreasing correlations are observed at the beginning of the ESDC, while the most stable periods for all currencies are the first and second phase of the GFC. The results provide several implications for central bank interventions and international portfolio diversification. The layout of the present paper is as follows. Section 2 presents the multivariate FIAPARCH-DCC model and the identification of the length and the phases of the two crises. Section 3 provides the data and a preliminary analysis. The empirical results are displayed, analyzed and discussed in Section 4, while Section 5 reports the concluding remarks.
نتیجه گیری انگلیسی
This paper examines the time-varying linkages among daily US dollar exchange rates expressed in five major international currencies, namely: EUR, JPY, CHF, GBP and AUD. Specifically, we employ a multivariate FIAPARCH(1,d,1)-DCC model, during the period from 2004 to 2011, focusing on the estimated dynamic conditional correlations among the currencies during different phases of GFC and ESDC. This procedure overcomes several limitations of the existing literature, since it investigates the second order moments dynamics of exchange rates taking into account long memory behavior, asymmetries and leverage effects. The current literature provides evidence on exchange rates co-movements and volatility spillover effects among several currencies. However, the empirical analysis either focuses on earlier periods which include the introduction of euro (e.g., Nikkinen et al., 2006, Perez-Rodriguez, 2006, Inagaki, 2007 and Antonakakis, 2012) or use high-frequency data to capture intraday volatility (e.g., Engle et al., 1990 and Kitamura, 2010) and the effects of political and risk news including earlier financial crises (Renaldo and Soderlind, 2010). On the contrary, our empirical investigation provides additional insights and several interesting findings, which contribute to both the crises literature and studies on exchange rates markets, since it analyzes the dynamic correlation patterns across the two recent financial crises. The majority of the conditional correlations among currencies decline across the different phases of the two crises, indicating their different vulnerability. The first and the second phase of GFC exhibit the most statistically insignificant correlation coefficients for all currencies, indicating that at beginning of the crisis news may be considered as a single-country case and the GFC signal has not been fully recognized. Moreover, the British pound shows the greater exposure at the beginning of the GFC, while the JPY seems to be a safe heaven currency during the second phase of the crisis. On the contrary, the only positive correlation change is observed for the pair of GBP-AUD during the third phase of the GFC, implying a currency contagion. At the fourth phase of recovery, currencies have different degrees of appreciation, indicating the end of the GFC. Across all phases of the two crises, the most cases of decreasing correlations are observed at the beginning of the ESDC. During the first phase of the Eurozone crisis, JPY, GBP and CHF have significantly negative correlations with EUR, indicating the uncertainty about the future of the Eurozone currency. At the last phase, only JPY and CHF continued to have the aforementioned behavior, implying that these currencies provided hedging benefits and could be characterized as safe heaven currencies. The results lead to important implications from investors’ and policy makers’ perspective. The decrease of exchange rates linkages during turmoil periods shows the different vulnerability of the currencies and implies an increase of portfolio diversification benefits, since holding a portfolio with diverse currencies is less subject to systematic risk. Moreover, this correlations’ behavior may be considered as evidence of non-cooperative monetary policies around the world and highlight the need for some form of policy coordination among central banks. Finally, the different patterns of dynamic linkages among major world currencies influence transnational trade flows and the activities of multinational corporations, as they create uncertainty with regard to exports and imports.