Information Transmission between Stock Markets in Hong Kong, Europe and the US : New Evidence on Time- and State-Dependence

This article performs a long-term investigation of information transmission between stock markets in Hong Kong, Europe and the US. The particular focus is on the time- and state-dependence of return spillovers and autocorrelations as well as the related potential deviations from informational efficiency. We use high-frequency data for the Hang Seng, the Euro Stoxx 50 and the S&P 500 index from 2000 to 2011 and conduct Granger causality inference based upon non-overlapping intraday returns. Results from structural break tests suggest that the process of information transmission is structurally stable over time. Moving window regressions, however, reveal short-lived temporary deviations from informational efficiency in the form of weak, but significant spillovers and return autocorrelations. Most pronounced are temporary negative spillovers from the US to Hong Kong as well as temporary positive spillovers from Europe to the US. Threshold model estimations finally indicate that spillovers to the European and the US market are only significant in the state of high realized volatility. Spillovers to Hong Kong, however, tend to be significant in the state of low realized volatility