Conference Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 279 - Temporal and Spatial Models in Business and Economics
Type: Contributed
Date/Time: Tuesday, August 9, 2022 : 10:30 AM to 12:20 PM
Sponsor: Business and Economic Statistics Section
Abstract #323491
Title: Mixed Models for Calendar Effects in Seasonal Adjustment
Author(s): Steven Mark Mance*
Companies: Bureau of Labor Statistics
Keywords: calendar effects; mixed model; seasonal adjustment; Fay-Herriot model; Current Employment Statistics
Abstract:

Economic time series may exhibit variation due to aspects of the calendar such as moving holidays, the number of business days in the month, or the length of time between survey reference periods. Statistical agencies often seek to remove these “calendar effects” during seasonal adjustment so as not to obscure economic fluctuations that are of more interest to data users. Calendar effects are typically estimated using a univariate time series regression with ARIMA errors (RegARIMA), but these estimates may be unreliable in short time series or when the model is misspecified (e.g., outliers are not accounted for properly). This paper presents a mixed model for estimating calendar effects that “pools strength” across many related series (in this case, employment in the same industry across all states). The mixed model is assessed by comparing out-of-sample forecast performance against the standard approach.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2022 program