Seasonal Differences in Employment between Survey and Administrative Data

Jeffrey A. Groen


This paper examines seasonal differences in monthly employment figures gathered from two Bureau of Labor Statistics programs. One is the Quarterly Census of Employment and Wages (QCEW), which is based on mandatory quarterly Unemployment Insurance reports; the other is the Current Employment Statistics (CES) survey. Despite using similar definitions of employment, QCEW and CES estimates are often different at micro and aggregate levels, both at a point in time and in seasonal patterns. At the aggregate level, the largest differences in growth rates between QCEW and CES employment occur from November and January. Three-fourths of the differences in monthly employment growth between QCEW and CES is due to reporting differences. Analysis of two matched samples of QCEW-CES micro data reveals that seasonal differences are related to imputation in the QCEW, the number and frequency of payrolls, and differences in the procedures used by establishments to compile QCEW and CES data.