Cyclical Behavior Analysis of Indian Market Using H-P Filter and Spectral Techniques.

Time Series (TS) data often exhibits a cyclical pattern. Security indices time series data is not an exception to that. Three forms of Indian security indices data are studied here for characterization of business cycles. These are of values, returns, and moving variance series of selected sectorial National Stock Exchange (NSE) indices together with the gross indices Nifty and Sensex by considering daily movements. The datasets are analyzed by Fast Fourier Transform (FFT), Short Time Fourier Transform (STFT) and wavelet-based methods with and without Hodrick and Prescott (H-P) filtering as pre-processing. The FFT analysis shows that most of the indices show business cycles of approximately quarterly duration, whereas wavelet studies identified some structural breaks in some of the considered datasets. [ABSTRACT FROM AUTHOR]