Trump Fires Labor Chief Over Weak Jobs Numbers

Government statistics agencies are historically insulated from politics, so they can accomplish the deeply technical task of tallying up activity in a $30 trillion, 160 million-job economy.

BLS is considered the gold standard among international labor data collectors. Economists consider its revisions process key to transparency and thoroughness, even if it sometimes makes BLS look bad when preliminary numbers turn out to be wildly off.

The establishment survey, on which the job growth numbers are based, comes from monthly contact with 121,000 employers nationwide, asking them how many workers are on their payrolls, how many hours they worked, and so forth.

The agency revises the numbers after evaluating information that wasn't available at the time of the initial report.

That includes information from businesses that don't respond before the report is released. The BLS continues to take reports from those businesses, which inform the revisions published in the following two months.

The unemployment rate, however, is not revised monthly.

The revisions, which have been conducted since 1979, can be so large that they change the perspective of the current state of economy, especially if a large company is late to report payrolls data.

The Labor Department also recalculates seasonal factors, a statistical quirk that aims to remove the seasonal patterns — weather, holidays school schedules and the like — that can influence employment trends.

The jobs numbers are also benchmarked yearly against unemployment insurance filings, to ensure that the survey data lines up with actual employment as reflected in administrative records.