Calendar effects in JDemetra+

The following description of the calendar effects in JDemetra+ is strictly based on PALATE, J. (2014).

A natural way for modelling calendar effects consists of distributing the days of each period into different groups. The regression variable corresponding to a type of day (a group) is simply defined by the number of days it contains for each period. Usual classifications are:

• Trading days (7 groups): each day of the week defines a group (Mondays,...,Sundays);

• Working days (2 groups): week days and weekends.

The definition of a group could involve partial days. For instance, we could consider that one half of Saturdays belong to week days and the second half to weekends.

Usually, specific holidays are handled as Sundays and they are included in the group corresponding to "non-working days". This approach assumes that the economic activity on national holidays is the same (or very close to) the level of activity that is typical for Sundays. Alternatively, specific holidays can be considered separately, e.g. by the specification that divided days into three groups:

• Working days (Mondays to Fridays, except for specific holidays),

• Non-working days (Saturdays and Sundays, except for specific holidays),

• Specific holidays.