CAUSES OF SEASONALITY

It is a very well-known fact that many economic series display seasonality; that is, they have an observable component consisting of a fairly constant shape repeated every 12 months. This component is often treated as being so easily explained that neither an exact definition nor an explanation of its origins is required. It is the objective of this paper to suggest that ignoring consideration of causation can lead to imprecise or improper definitions of seasonality and consequently to misunderstanding of why series require seasonal adjustment, to improper criteria for a good method of adjustment and to have implications for the evaluation of the effects of adjustment both on a single series and when relating two or more series. These considerations do not necessarily lead to better practical methods of adjustment, but they should lead to a better understanding of how to interpret time series and econometric analysis involving seasonal components and seasonally adjusted series. The only other author, prior to this conference, who emphasizes causation of seasonals appears to be BarOn (1973).

There are at least four, not totally distinct, classes of causes of seasonal fluctuations in economic data. These classes are discussed in the following sections.

Calendar

The timing of certain public holidays, such as Christmas and Easter, clearly affects some series, particularly those related to production. Many series are recorded over calendar months, and, as the number of working days varies considerably from one month to another in a predetermined way, this will cause a seasonal movement in flow variables, such as imports or production.