Modeling Recessing Effects and the Consequences on Seasonal Adjustment

• Activity Number: 523
• Type: Topic Contributed
• Date/Time: Wednesday, August 7, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): Demetra Lytras*+ and William R. Bell
• Companies: US Census Bureau and U.S. Census Bureau
• Keywords: ARIMA model ; intervention ; outlier ; regression ; time series

Abstract

The 2008 U.S. recession generated many questions as to whether the large declines observed in many time series affected their seasonal adjustments. We investigated this by comparing seasonal adjustments obtained with versus without adjustments for recession effects for 23 monthly U.S. Census Bureau economic time series. We produced the recession adjustments using intervention models that fit short series of level shifts to series with relatively short (less than five month) declines, and ramps to series with longer declines. In all cases, adding the intervention terms substantially improved model fit; however, the effect of the recession preadjustments on the seasonal adjustments was generally small. The Census Bureau has started including these recession adjustments to series showing pronounced recession effects.