Volatility Spillover Effect: A Semiparametric Analysis of Non-Cointegrated Process

Stock market volatility is highly persistent and exhibits large fluctuations so that it is likely to be an integrated or a near integrated process. Stock markets' volatilities from different countries are intercorrelated, but are generally not cointegrated as many other (domestic) factors also affect volatility. In this paper, we use a semiparametric varying coefficient model to examine stock market volatility spillover effects. Using the estimation method proposed by Sun et al. (2011), we study the U.S./U.K. and U.S./Canadian stock market volatility spillover effects. We find striking similar patterns in both the U.S./U.K. and the U.S./Canadian markets. The stock market volatility spillover effects are strengthened when the currency markets experience high movement, and the spillover effects are asymmetric depending on whether a foreign currency is appreciating or depreciating.