An Appraisal of Multivariate Seasonal Adjustment

• Activity Number: 119
• Type: Topic Contributed
• Date/Time: Monday, August 5, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Business and Economic Statistics Section
• Author(s): Tucker S. McElroy*+
• Companies: U.S. Census Bureau
• Keywords: Signal extraction ; Dynamic factor models ; Vector time series ; Trends ; Seasonality

Abstract

There has been persistent interest among economists for statistical agencies to research and implement a multivariate seasonal adjustment procedure. Recent advances in vector time series modeling -- including dynamic factor models and vector structural time series models -- make such a project possible. Our research first examines the dimension reduction techniques available, and how the resulting lower-dimensional factor time series can be jointly modeled. Several approaches are developed and discussed, allowing for an X-11 type of treatment, based upon forecast extension with a fixed moving average filter, as well as a fully model-based treatment. Practical issues, such as model identification and estimation, are also discussed, and the proposed methods are tested upon Census Bureau retail, manufacturing, and construction series.