On fractional filtering versus conventional filtering in economics

In this study, we compare the Hodrick-Prescott Filter technique concerning the Fractional filtering technique, which has recently started to be used in various applied sciences, i.e., physics, engineering, and biology. We apply these filtering techniques to the quarterly GDP data of Turkey, which span the period 1988:1 2003:2. The estimated filtered series are then compared using classical statistical tool MSE (Minimum Square Error) and with real-life evidence such as crisis periods, recessionary, or boom periods. In the second part of the study, we use generated data that exhibits the essential characteristics of economic data to see the effects of filtering on these data and trace the effects of these filtering’s on decomposed series.