An important practical problem for federal egencies that publish economic data is the seasonal adjustment of mixed frequency stock and flow time series. This may arise in practice due to changes in funding of a particular survey; for example, a series published at a quarterly frequency in the past might transition to a monthly publication schedule. We demonstrate that the X-12-ARIMA method can be applied to such mixed frequency data by deriving formulas for conditional expectations that can be applied to forecasting problems. Moreover, standard errors are easily produced for this method as a by-product of the calculations. The efficiency of the parameter estimation procedure is investigated through simulation, and the seasonal adjustment method is demonstrated on several mixed frequency time series.
Keywords: Seasonality; Trends; Signal extraction; Forecasting
Biography: Tucker McElroy is a mathematican at the U.S. Census Bureau, whose research interests include seasonal adjustment, time series model specification tests, signal extraction, forecasting, nonparametric time series methods, long memory time series modeling, and extreme value estimation.