|
Activity Number:
|
205
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Business and Economic Statistics Section
|
| Abstract - #304113 |
|
Title:
|
Issues in Trend-Cycle Estimates for Official Statistics
|
|
Author(s):
|
Begoña Martín*+ and Duncan Elliott and Peter Meszaros and Craig McLaren
|
|
Companies:
|
Office for National Statistics and Office for National Statistics and Statistics Canada and UK Office for National Statistics
|
|
Address:
|
Cardiff Road, Newport, NP10 8XG, United Kingdom
|
|
Keywords:
|
Trend estimation ; Henderson filter ; ARIMA ; false turning points
|
|
Abstract:
|
The current economic climate emphasizes the importance of identifying turning points in time series quickly and accurately. Trends are often used to describe underlying movements in series, so need to perform well in periods of changes to avoid misleading users. We examine the properties of short-term trend estimation methods with regards to accuracy, bias and revisions. We evaluate the performance of various filters - including the recent Cascade Linear Filter (Dagum and Luati, 2009) - and discuss the underlying motivation behind these measures.
|
