Modeling of BLS and Census Bureau Seasonal Time Series with Frequency Specific Generalized Airline Models With GenAirNBB

David Findley, Thomas D. Evans, Jean Palate, and Richard B. Tiller


We present results from a study of the applicability of the Frequency Specific Models (FSMs) of Aston, Findley, Wills and, Martin (2004) to a broad range of seasonal time series, from the Bureau of Labor Statistics and from the Census Bureau. These models are generalizations of the most widely used seasonal ARIMA model, the (0,1,1)(0,1,1) or "airline" model. They make possible improved modeling of series whose seasonal movements are dominated by frequency components with frequencies in a subset of the seasonal frequencies 1, 2, 3, 4, 5 and 6 cycles per year (for monthly data). The study was done with a versatile menu-driven program, named GenairNBB, developed by the National Bank of Belgium. The software has batch options for processing large numbers of series that will be demonstrated. It will be available from the web site of the bank.