Resumen de A note on the use of calendar regressors

Leandro Navarro Pablo

• In this paper I analyse the implications of using three different variables aiming at capturing the “trading day” effect. Specifically, the proposal outlined in Bógalo and Quilis (2006), which is consistent with Eurostat’s (2007 and 2002), Cleveland and Devlin (1982), and the one used in the programs Tramo/Seats. It is found that Eurostat’s (2002 and 2007) proposal presents some drawbacks with respect to the standard regressors used in Tramo/Seats and when compared with the variables brought forward by Cleveland and Devlin (1982).

  Moreover, although these variables might show some downsides, they offer the best results of all the methods considered.