Realized correlation of high-frequency data based fourier analysis
Financial high-frequency data are not equally spaced, standard correlation measures such as Realized Correlation can not be directly applied to the raw data which have to be homogenized by interpolating. Fourier-Realized Correlation is a method which uses all the tick-by-tick data with no need to change their structure. In this paper, Fourier-Realized Correlation of stock returns, the efficiencies of the two methods are compared. Empirical results show that Fourier-Realized Correlation is a better choice in terms of stability.