Volatility Spillover Effect: A Semiparametric Analysis of Non-Cointegrated Process
用半参数变系数模型研究美国与英国、加拿大之间的股市波动溢出效应，发现汇率波动大时溢出效应增强，且存在不对称性。
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 Sun , Y. , Hsiao , C. , Li , Q. ( 2011 ). Measuring correlations of integrated but not cointegrated variables–a semiparametric approach . Journal of Econometrics 164 : 252 – 267 .[Crossref], [Web of Science ®] , [Google Scholar]), 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.