The Effects of Seasonal Adjustment Methods in Nonparametric Trend-Cycle Prediction

• Activity Number: 490
• Type: Topic Contributed
• Date/Time: Wednesday, August 6, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): Estella Dagum*+ and Silvia Bianconcini
• Companies: University of Bologna and University of Bologna
• Keywords: Asymmetric filters ; turning point detection ; reproducing kernel Hilbert space ; X12ARIMA ; TRAMO-SEATS

Abstract

In a previous study, Dagum and Bianconcini (2013) have shown that using local time-varying bandwidth parameters in asymmetric nonparametric trend-cycle filters reduces significantly revisions and turning point detection respect to the Musgrave (1964) standard approach. The authors observed that the best choice of local time-varying bandwidth is the one obtained by minimizing the distance between the gain functions of the asymmetric and the symmetric filter to which it must converge. The purpose of this study is to evaluate the effects of the seasonal adjustment method when the real time trend is predicted with such nonparametric filters. The seasonal adjustments are made with the X12ARIMA and TRAMO-SEATS on a large sample of series.