About the Early Benchmarks
The Federal Reserve Bank of Philadelphia began releasing quarterly early benchmarks of state employment estimates in November 2022, accompanied by an article that explains the purpose and methodology and discusses how to interpret the analysis.
The primary purpose of the early benchmark analysis is to produce timely estimates of state payroll jobs that closely predict the annual benchmark revisions released by the Bureau of Labor Statistics (BLS) each March. Our state estimates have performed well in the past, despite our researchers using some simplifying procedures and having access to less complete data than does the BLS.
A secondary purpose of the Philadelphia Fed’s early benchmarks is to augment our understanding of national economic trends by searching for signals gleaned from the revisions to state estimates. Revisions that are:
- significant in magnitude;
- pervasive across a majority of large states; and
- persistent over multiple quarters
may suggest the direction in which national employment data will be revised.
In our experience, relatively small overall revisions suggest that the national estimate will not be significantly revised even if the revisions are pervasive and persistent. Relatively large overall revisions over three or more quarters suggest that the national estimate may be revised slightly in the same direction as the overall state revisions. As our article stated, “We need to track this work over more years to learn whether our early benchmarks regularly predict the direction of data revisions to the CES estimates of national data.”
For our analytical purposes, we compute the sum of our early benchmark state estimates solely as a metric to assess the first two criteria across multiple quarters and across years. We compute these estimates for cumulative quarters as data become available in order to assess the third criteria. The estimate of U.S. employment obtained from the sum of our early benchmark state estimates is not designed nor intended to be an accurate measure of national employment. Moreover, the BLS routinely warns that because of statistical limitations, it “does not compile a ‘sum-of-states’ employment series and cautions users that such a series is subject to a relatively large and volatile error structure.” This caveat also applies to our early benchmark series.