The nation’s official employment numbers come from the Bureau of Labor Statistics, based on a monthly survey of businesses and a survey of households. The BLS numbers get annually adjusted based on other data sources and refined estimates. Some states produce numbers as well. Some private firms that service large numbers of businesses for payroll or human resources needs also estimate employment and job changes. And then there is the Philadelphia branch of the Federal Reserve Bank. All the Federal Reserve branches have excellent economists who turn out good research. The Philadelphia branch specializes in the employment numbers, in particular making an estimate of what the BLS revisions will look like.
It’s review of state unemployment records indicated that for the period from March through June, the initial number released by the Bureau of Labor Statistics were likely erroneous for 27 states. And the analysis suggested that to a degree of statistical certainty, actual employment was lower in 27 states and higher in two. Whereas the BLS numbers look like over 1% employment growth, the estimate of the benchmark revision shows only about one-tenth of a percent. Oddly enough, Minnesota is one of those two states whose jobs might be revised higher. Which seems unlikely based on other data. (PFR Report)