|
Abstract:
|
Traditionally, seasonal adjustment also involves the elimination of average calendar influences. By contrast, no separate adjustment usually takes place for bridging days, school holidays and weather-induced effects. These effects are investigated in the paper based on RegARIMA models. It will be shown that the three variables have a significant influence on the German production index. However, problems are apparent, these being that the adjusted results are not always plausible, the magnitude of the influences may depend on the business cycle and catch-up effects are probable but impossible to quantify. Hence, it is argued that the decision about which variables should be used for the calendar adjustment of officially published data should take into account more criteria than only statistical significance.
|
