Model-Assisted Estimation of Unemployment Rates for Longitudinal Surveys with an Application in the Current Population Survey

• Activity Number: 653
• Type: Contributed
• Date/Time: Thursday, August 7, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Survey Research Methods Section
• Author(s): Daniel Bonnery*+ and Yang Cheng and Partha Lahiri
• Companies: Joint Program in Survey Methodology/U.S. Census Bureau and U.S. Census Bureau and University of Maryland
• Keywords: Longitudinal surveys ; Unemployment rate ; Model assisted estimation ; Current Population Survey

Abstract

The Current Population Survey (CPS), sponsored jointly by the U.S. Census Bureau and the U.S. Bureau of Labor Statistics (BLS), is the primary source of labor force statistics for the population of the United States. The official unemployment rate estimates are based on a composite estimator, known as the A-K estimator, which borrows strength from past survey data. We extend the standard model-assisted method with estimated total for the auxiliary variables using models that are appropriate for longitudinal surveys. We critically examine theoretical properties of the proposed method. Using a Monte Carlo simulation experiment, we study different models for employment evolutions and study simulated properties of different estimators. We also compare these estimators when applied to the CPS data.