JOLTS Reported Separations, Item Imputation, and Sample Rotation: Impact on JOLTS-CES Divergence

• Activity Number: 89
• Type: Contributed
• Date/Time: Sunday, August 4, 2013 : 4:00 PM to 5:50 PM
• Sponsor: Government Statistics Section
• Author(s): Mark Crankshaw*+
• Companies: Bureau of Labor Statistics
• Keywords: item imputation ; synthetic estimates

Abstract

The Bureau of Labor Statistic's Job Opening and Labor Turnover survey (JOLTS) attempts to measure US labor market dynamics on a monthly basis. The JOLTS survey estimates both the flow of employment into business establishments (hires) and the flow of employment out of business establishments (separations). In theory, it is assumed that the levels of these two types of flow should relate directly and proportionately to the trend in the overall employment level. A number of factors prevent the employment trend and the net flow of hires and separations from equilibrating on a monthly basis. However, it is expected in theory that over time this equilibration should eventually occur.

We will present evidence that even over time, for reported JOLTS data, the net flow of hires minus separations greatly exceeds reported employment trend and that this disequilibrium is the result of a significant deficiency in reported separations or excess of reported hires. Finally, we will present the corrective action adopted by JOLTS to account for the divergence between Current Employment Survey (CES) employment and JOLTS hires minus separations.