|
Activity Number:
|
232
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, August 8, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Business and Economics Statistics Section
|
| Abstract - #306884 |
|
Title:
|
Exploring Model-Based Seasonal Adjustment for Industry Employment Statistics
|
|
Author(s):
|
Christopher Manning*+ and Jeffrey A. Smith
|
|
Companies:
|
Bureau of Labor Statistics and Bureau of Labor Statistics
|
|
Address:
|
2 Massachusettes Ave., NE, Washington, DC, 20212,
|
|
Keywords:
|
seasonal adjustment ; X-12-ARIMA ; TRAMO/SEATS ; X-12-SEATS
|
|
Abstract:
|
This study explores the use of ARIMA model-based seasonal adjustment in the Current Employment Statistics (CES) program and the implications it might have on seasonally adjusted CES series. Theoretically, ARIMA model-based seasonal adjustment has the advantage over the X-11 method by shaping its seasonal decomposition to the individual series while the X-11 method must choose from a finite number of built-in filters. We attempt to determine whether this theoretical advantage leads to actual benefits by examining CES employment estimates seasonally adjusted under the different methodologies. Our test will use X-12-ARIMA, TRAMO/SEATS, and X-12-SEATS. TRAMO/SEATS and X-12-SEATS are expected to differ mainly in model selection. The evaluation of the seasonally adjusted series will check for residual seasonality and compare stability diagnostics and smoothness statistics.
|
