By Luca Benati
In recent years, band-pass filtering—the non-structural, frequency-domain based decomposition of economic time series into trend and cyclical components—has become more and more popular among macroeconomists, as a way of capturing and describing business cycle stylised facts. Compared with the Hodrick-Prescott filter, the band-pass filter offers the advantage of allowing the researcher to target a specific frequency band, thus extracting from the series of interest all the components associated with that band, while essentially discarding all the others. The growing interest of the macroeconomics profession in band-pass filtering techniques is demonstrated, first by the number of recent papers on business cycle stylised facts that make use of the band-pass filter, second, by the inclusion in the recent Handbook of Macroeconomics of a chapter on US post-World War II business cycle stylised facts entirely based on band-pass filtering, and third, by the continuing attempts to develop new and better approximations to the ideal band-pass filter.
Band-pass filtering, cointegration, and business cycle analysis