We find evidence of a cyclical and seasonal bias in the annual benchmark revisions to the monthly level of non-farm payroll employment. We propose a change to the methodology that would remove the seasonal bias and ensure that the benchmark process does not artificially influence the estimated...