|
Activity Number:
|
181
- Contributed Poster Presentations: Government Statistics Section
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 31, 2017 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Government Statistics Section
|
|
Abstract #322927
|
|
|
Title:
|
Time Series Reconciliation: a Least-Squares Solution for the Case of High-Frequency Series Exceeding the Period Covered by the Low-Frequency Series
|
|
Author(s):
|
Luis Frank*
|
|
Companies:
|
|
|
Keywords:
|
Time series reconciliation ;
economic time series ;
official statistics ;
least-squares criterion
|
|
Abstract:
|
The paper presents an alternative method - based on the least squares criterion - to reconciliate time series observed at different frequencies. The main advantage of this method - besides its simplicity to reconciliate one or more high frequency (e.g. monthly) series with a low frequency (e.g. quarterly) series - is that it enables the extrapolation of the reconciled monthly time series beyond the last available quarter, avoiding the usual practice of forecasting an additional quarter to achieve a soft reconciliation with the last months of the monthly series. It is worth noting however that the ``reconciled' monthly values obtained with this procedure will not add up (or average) exactly to the corresponding values of the quarterly time series. The paper also compares the new procedure with other methods widely spread in official statistical bureaus using real data from the National Accounts of Argentina.
|
Authors who are presenting talks have a * after their name.
