Quarterly Benchmarking for the Current Employment Survey

Matthew Dey and Mark A. Loewenstein

Abstract

This paper proposes a quarterly benchmarking procedure for the Current Employment Survey (CES) that explicitly corrects for differences in seasonality between the CES and the Quarterly Census of Employment and Wages (QCEW). Abstracting from seasonal adjustment issues, we show analytically that the proposed procedure yields improved estimates of March employment. This is in spite of the fact that revisions in various quarters are likely to be opposite signed. More frequent benchmarking is especially advantageous when CES errors are correlated over time, as tends to happen at turning points in the business cycle. The March revision is generally not a good measure of monthly errors in the CES. A small March revision does not in and of itself imply small monthly errors. The performance of the proposed quarterly benchmarking procedure depends on the variance of the CES estimator, potential errors in the QCEW, and on how well the seasonal factors in the CES and QCEW are estimated. Simulations show that the variance of the proposed estimator is smaller than that of the CES under the plausible assumption that errors in the QCEW are relatively small compared to errors in the CES. The larger the errors in the CES, the better is the relative performance of the proposed quarterly benchmark estimator even accounting for the fact that seasonal factors will be estimated less precisely.