|
Activity Number:
|
26
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 8, 2004 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #300735 |
|
Title:
|
A Method of Reconciling Discrepant National and Subnational Employment and Unemployment Estimates
|
|
Author(s):
|
Swamy A.V.B. Paravastu*+ and Jatinder S. Mehta and i-Lok Chang
|
|
Companies:
|
Bureau of Labor Statistics and Temple University and American University
|
|
Address:
|
Room 3885, Postal Square Building , Washington, DC, 20212,
|
|
Keywords:
|
small-area estimation ; geographical variation ; nonlinear regressions ; cross-sectional heteroscedasticity ; iteratively rescaled generalized least squares ; errors-in-variables
|
|
Abstract:
|
The principal goal of this paper is to improve the estimation methods adopted in the Local Area Unemployment Statistics (LAUS) program within the Bureau of Labor Statistics (BLS). To jointly model all the available data on employment and unemployment for small areas, the estimators applied to these data are considered. Even though some of these estimators are unbiased in probability sampling, the estimates of employment or unemployment provided by these estimators are different, containing different magnitudes of nonresponse and measurement-error biases and nonsampling and sampling errors. This paper develops a method of estimating these biases and errors. It takes a pair of estimates of employment or unemployment for each of several geographical areas and finds a good model of their conditional variations across areas both at a point in time and through time. This model improves one of the pair of estimates and corrects the other estimate for nonresponse and measurement-error biases and for sampling and nonsampling errors. The improved estimate is equal to the corrected estimate. The method's practical behavior is demonstrated on a real dataset.
|
