Time and frequency domain in the business cycle structure

The paper deals with identification of cyclical behaviour of business cycle from time and frequency domain perspectives. Herewith, commonly used methods for obtaining growth business cycle are investigated - the first order difference, the unobserved component models, regression curves and filtration using Baxter-King and Christiano-Fitzgerald band-pass filters as well as Hodrick-Prescott high-pass filter. In the case of time domain analysis identification of cycle lengths is based on dating process of the growth business cycle. For this reason, methods such right and left variant of naive techniques as well as Bry-Boschan algorithm are applied. In the case frequency domain analysis of cyclical structure trough spectrum estimate via periodogram and autoregressive process with optimum lag are suggested. Results from both domain approaches are compared. On their bases recommendation for cyclical structure identification of growth business cycle of the transition economy type (the Czech Republic) are formulated. In the context of the time domain analysis evaluation of unity results of de-trending techniques from identification turning point points of view is attached. All analyses are done on the quarterly data of the gross domestic product, the total industry excluding construction, the gross capital formation in the period 1996/Q1-2008/Q4 and on the final consumption expenditure in the period 1995/Q1-2008/Q4.