Reconciling Employment Differences between Administrative and Survey Data

This paper investigates differences in employment figures gathered from unemployment-insurance tax filings under the Quarterly Census of Employment and Wages (QCEW) and employment figures reported through the Current Employment Statistics (CES) survey. Since these two Bureau of Labor Statistics (BLS) programs both collect monthly employment from an establishment for the same reference period, the employment figures should generally be identical. However, differences exist at the micro and aggregate levels, both at a point in time and in seasonal patterns. We analyze employment differences for 200 establishments with large differences in 2005-06 and determine the scope of and patterns in the differences. We then report on findings from a response analysis survey (RAS) in which the establishments were asked about reasons for differences in employment reported on the tax and survey forms.