The analysis of seasonality in economic statistics: a survey of recent developments

Abstract

This article describes the EUROSTAT activities in the field of seasonal adjustment and trend extraction in economic time series. They follow a working program which has been set up during 1995. The attention focuses on X12-REGARIMA (X12 in short), a last update of the X11-family from the Bureau of the Census (see Findley and al., 1996), and on SEATS-TRAMO (see Gomez and Maravall, 1996) which implements the ARIMA-model-based approach to decompose time series. Three main directions are currently followed: evaluation and comparison of these two models, construction of a software embodying and interfacing X12 and SEATS-TRAMO, and training in applied time series analysis. The preliminary results which have been obtained on the difficult task of comparing both methods are discussed and the design of the software in construction is presented.