Since July 2003, the Banque de France has been using seasonally adjusted (SA) data for the monthly reporting of national monetary developments, with renewed statistical tools. Before the start of the single currency in 1999, the Banque de France already calculated seasonally adjusted data, using a rather old method, X11-ARIMA. Due to the shortcomings of this means, the Banque de France has developed a new method of seasonal adjustment, using both TRAMO-SEATS and X12-ARIMA, and defining a specific revision policy for each SA series. his paper aims at presenting and explaining the choices made by the Banque de France regarding the implementation of the new production process of seasonally monetary and loans series. In the meantime, the theoretical background related to the concept of seasonality and to various seasonal adjustment methods is shown in order to throw light on these choices, thus not from a research angle. The paper firstly provides information about the concept of seasonality as well as the existing methods of seasonal adjustment. The new production process of SA monetary data at the Banque de France is then described. Two examples related to the seasonal adjustment of loans to enterprises and to housing loans are included in order to stress the difficulties implied by the monthly production of SA data, as well as the impact on then output of the choices made by the producer regarding the modelling of the seasonality.
Classification JEL : C22, C51
Keywords : Seasonal adjustment methods, monetary aggregates, outliers, SARIMA models.
Updated on: 06/12/2018 10:59