ABSTRACT

Economic time series are typically available at the monthly frequency of observations. A key feature is the presence of seasonality and calendar effects, which account for much of the variation in the series. Modeling and extracting these components have thus constituted an important problem in the analysis of economic time series; see Zellner (1978, 1983), Nerlove et al. (1979), Hylleberg (1992), Pen˜a et al. (2001), and Ghysels and Osborn (2001); Findley (2005) discusses some recent advances in seasonal adjustment.