Ten lectures on wavelets
An introduction to wavelets
| Hi-index | 0.00 |
This paper investigates the causal links between the world's largest stock markets, namely the U.S. market, the U.K. market, and the Japanese market, over various time horizons. The major innovation of this paper is to apply the wavelet multiscaling method into the study of the international stock market linkage. The main empirical results from the wavelet multiscaling method support three conclusions. First, there is a significant bi-directional causal effect among the three markets when viewed in terms of high frequency behavior. Second, no consistent causality effects are observed in the frequency of 16 days, which we can view approximately as monthly data, between all three major stock markets. This empirical finding implies that the three major stock markets are strongly connected in their daily or weekly returns, while in the monthly returns level, there does not exist a consistent causality effect. Third, the empirical evidence we provide supports the findings of Ramsey & Lampart (1998) and Gencay et. al. (2002) that there is not one global causality relation prevailing over all time scales.