The effect of financial data noise on the long-term co-movement of stock markets

Due to advances in information technology and development in economies, the linkage among international markets becomes more significant especially for portfolio managers. Meanwhile, the noise component of time series data observes to be imperfect in doing financial analysis such that testing the theories becomes difficult. Applying the wavelet de-nosing method to the co-integration model, this paper investigate the effect of financial time series noise on long term behavior of 16 capital market indices. The weekly closing index price of the selected markets is used and findings revealed that the de-noised time series are more co-integrated compared to the noisy data. Moreover, using de-noised time series would give profound view in the long-term co-movement analysis.

Keywords:

financial data noise , wavelet de-noising , thresholding function , long-term co-movement