|
Abstract:
|
Most of economic indicators are linked to a daily activity which is usually summed up each month or each quarter. In this case, the number of working days, which varies from one month to another in a quasi-predetermined way, can explain some short-term movements in the time series. Due to the regularity of the calendar, this Trading-Day effect (TD) corresponds to specific frequencies and could be detected in the spectrum of the series. X-12-Arima implements such a "visual spectral test." In a first part, the "theoretical" TD frequencies are derived from the spectrum of the Gregorian calendar. But a real TD effect, a linear combination of these basic frequencies, can correspond to different frequencies. In a second part, the study of several thousands of real time series permits to exhibit the "real" TD frequencies, those one should use in a spectral test to detect a trading-day effect.
|
