Currency Risk and Volatility Spillover in Emerging Foreign Exchange Markets

1
Electronic copy available at: /abstract=1650049
In this paper, potential cross-country spillovers from both major stock markets and major foreign exchange markets to emerging foreign exchange markets in Asia and Latin America are investigated. The methodology allows for both asymmetry in volatility transmission and the potential impact of risk from own volatility of returns. In the case of stock market spillovers, early studies examined interdependence in terms of the conditional mean returns but later work also considered the possibility that volatility might be transmitted across markets (Hamao, Masulis, and Ng, 1990, 1991; Ng, Chang and Chou, 1991; Lin, Engle and Ito, 1991; Worthington and Higgs, 2004; Lee and Stewart, 2009). In a further development, Koutmos and Booth (1995) demonstrated how potential asymmetries in the volatility spillover mechanism - the leverage effect - can be accommodated using a multivariate exponential GARCH model. Applying the model to data on stock prices in London, New York and Tokyo, they found strong evidence of asymmetric spillovers. Subsequently, In et. al. (2001) found some evidence of asymmetric volatility transmission involving the Hong Kong, Korea and Thailand stock markets during the 1997-98 Asian financial crisis. To take account of market risk, Elyasiani and Mansur (1998) employed a multivariate GARCH-M model in which the conditional variance of bank returns is a potential determinant of the mean. The model specification in this paper draws upon both Koutmos and Booth (1995) and Elyasiani and Mansur (1998). The data come from two major stock markets (the US SP500 and Japan Nikkei225) and two major foreign exchange markets (the euro and Japanese yen), and ten Asian and Latin American foreign exchange markets: Brazil, Chile, China, Mexico, Colombia, Peru, Indonesia Philippines, South Korea, and Thailand. All of the latter are classified as emerging markets by the Morgan Stanley Capital International (MSCI) Barra. With the exception of China, all of the emerging markets in the sample experienced a serious currency or financial crisis during 1980’s and 1990’s. As a consequence of the crises they adopted a floating or a managed float exchange rate regime. One might expect therefore that these foreign exchange markets would have become less restrictive and more cointegrated with major foreign markets (Fraser and Power, 1997; Brooks et al, 2000; Bekaert and Harvey, 2000; Koutmos and Saidi, 2001; Selcuk, 2005). Applying a modified multivariate EGARCH model, we find evidence of volatility spillovers from past returns and volatility of the major stock and foreign exchange markets to all emerging markets, except Columbia. On average, the spillovers are larger to Asia than Latin America. There is also evidence of significant leverage effects and volatility persistence and, in the cases of Chile and Peru, an impact of risk on returns was found. The remainder of the paper is organized as follows. Section 2 describes a modified multivariate EGARCH methodology. Section 3 presents the data and empirical results, and Section 4 concludes.
Abstract
This paper employs a modified multivariate GARCH model to test for cross-country mean and volatility transmission among ten emerging foreign exchange markets in Asia and Latin America, together with potential spillovers from major external stock and foreign currency markets. The framework allows for possible risk, leverage and persistence effects. The estimates suggest the presence of both regional spillovers and the transmission of shocks from external stock and foreign exchange markets. The major external influences were the Japanese yen and the US SP500. On average, the spillovers from external markets were larger to Asian than Latin American currency markets. On the other hand, regional co-movement was found to be higher in Latin America. There is evidence of significant leverage effects and volatility persistence and, in the cases of China and Peru, an impact of risk on returns was found. The findings also suggest that allowing for regional cross-market and external effects leads to an increase in the estimates of the risk factors.
Key Words: Volatility spilloverБайду номын сангаас Exchange rate, Asymmetric volatility, GARCH
1. Introduction
There is now a sizeable empirical literature devoted to establishing the extent to which financial shocks are transmitted across countries. Much of the work has focused on stock markets, where a number of studies have found evidence of both mean and volatility spillovers. The spillovers typically emanate from major markets, in particular the US (see, for example, Hamao et al., 1990; Theodossiou and Lee, 1993; Susmel and Engle, 1994; Koutmos and Booth, 1995). A related strand of the literature has examined within-country interdependencies between stock markets and foreign exchange markets. The message from these studies is that volatility tends to be transmitted from stock markets to foreign exchange markets but not vice versa. For example, Kanas (2006) examined data on six developed countries and found that volatility spillovers from stock market returns to exchange rates were significant for all countries except Germany, but there were no instances of significant spillovers in the reverse direction.