The Labor Department revised the number of jobs added in July up by 40,000 jobs and the number of jobs added in August up by 46,000. These revisions have some people asking questions about the way the Bureau of Labor Statistics measures employment.
Here's a breakdown: the numbers in the monthly jobs report come from two places the household survey and the establishment survey. The household survey is used to calculate the unemployment rate and the establishment survey is used to calculate how many jobs the economy gained or lost the previous month.
The household survey data comes from Census Bureau workers going to people's homes and asking them if they are working, if they are looking for work, etc. The establishment survey comes from the BLS contacting businesses and government agencies and asking them about the total number of employees they have, how many hours they've worked, etc. The BLS can't talk to every single person or every single business so they talk to about 140,000 businesses and government agencies and 60,000 households, and they use that data to calculate the job situation for the entire economy.
The revisions in the latest jobs report come from the establishment survey, and the reason they happen is that the BLS doesn't have all the information at the time the numbers are calculated. The BLS gives businesses and government agencies a deadline to submit their data but not everyone makes it. The information that comes in after the cut off date is what causes the revisions. BLS economist Megan Barker explained it to me this way:
After the press release, we are continuing to get more data so that can cause slight movements in what is showing as employment change over the month. We do have a second and a third revision to all of our data.
So if a business didn't get the BLS their data in time for the official release, they'll show up in the revision. Another reason the numbers change Barker says, it that sometimes the BLS will choose to hold data. If the numbers they get from a business seem unusually high or how, the BLS may wait until it can verify them.
The revisions in the jobs numbers can look big at times, but the BLS says you have to remember that they are counting a lot of jobs.
"You are forgetting that what we are really measuring is 133 million jobs and so in 133 million jobs the over the month change is not terribly large, and the revisions that you see are small adjustments to that over the month change," Karen Kosanovich, another economist at the BLS, told me.
The unemployment rate fell to 7.8 percent in September, that number comes from the household survey.
For more about the jobs numbers listen to our show, Episode 392: Keeping The Biggest Secret In The U.S. Economy.