The number of working days can explain some short-term movements in the time series. One more Saturday in a month for example drastically impacts the retail trade turnover in European countries. Apart the day composition of the month, other calendars effects such as public holidays or religious events may also affect the series. These periodic fluctuations, as well as the seasonality, are usually detected and eliminated in order to exhibit the irregular or non-periodic movements which are probably of most interest and importance.
In this presentation, we focus on two important practical problems: the design of adequate regressors taking into account the specificities of national calendars and the choice of a correct set of regressors, trading off between the parcimony of the model and the quality of the estimation.
Keywords: Trading-day adjustment; Seasonal adjustment; Reg-Arima models
Biography: Dominique Ladiray is currently in charge of the Short-Term Statistics Department at INSEE, the French National Institute for Statistics and Economic Studies. He is also a Professor at the French National School of Statistics (ENSAE), school from which he graduated in 1985. His research interests are in time series analysis with an emphasis on seasonal adjustment, short-term forecasting and official statistics; he has published papers and books on these topics. He has over twenty years experience in applied time series analysis. He has also conducted courses in seasonal adjustment in numerous countries. He can be reached at INSEE, 18 boulevard Adolphe Pinard, 75014 Paris, France. Office phone number is +33-141176781; email is firstname.lastname@example.org