January is the month for “benchmark revisions,” which is when BLS revises its sample-based estimates to match other comprehensive data sources. Preliminary estimates of the benchmark revision, released last fall, showed that with this benchmark revision, the over-the-year change in employment between March 2012 and March 2013 is likely to be revised downward by 124,000. This means that we are likely to find that the gap in the labor market due to the Great Recession and its aftermath is larger than we think it is. Right now, the gap is estimated to be 7.7 million, if the preliminary estimates hold, the gap is instead over 7.8 million.
Another thing happening with this benchmark revision is a code change that moves nearly half a million in employment from the Private Household industry—which is not counted in the establishment survey—to the Education and Health Care Services industry, which is. This means the level of employment in the establishment survey will increase to reflect this coding change. I checked with BLS and they reported they will be revising employment, hours, and wage data back to 1979 to reflect this change.
The household data (which is where the unemployment rate and the labor force participation rate come from) will also be affected by new population controls, but in accordance with standard practice, earlier household data will not be revised to reflect these changes. This means that if you want to know the change in unemployment and labor force participation between December and January, you can’t look at the regular published data, and instead must look at the extra information provided, which will show the over-the-month change after removing the effect of the new controls. Look to tables B and C here to see what information was provided last year at this time.