Imagine for a minute that I give you a jar of jellybeans and ask for your best guess at how many are in the jar. I give you ten seconds, and, in that time, you’re allowed to do whatever you want including opening the jar and taking some of the jellybeans out or perhaps even taking pictures of the jar.
After those ten seconds are up, I take the jar away and then give you one hour to come up with your best guess. You can use whatever statistical techniques you want, whatever tools you want, and you can bring in a whole team of experts to help you. After that hour, though, your guess is due. At the end, you give me a preliminary guess of 1,597 jellybeans in the jar.
I then give you the jar again with the same rules, except this time, you get five whole minutes to examine it and then I give you and your team ten hours to give me a new, revised guess. You come back and inform me that you would like to change your guess to 1,595 jellybeans.
In response, newspaper headlines across the country run the story that your revised guess was a “massive” and “unprecedented” divergence from your preliminary guess. You get accused of having an axe to grind against the jellybean industry and the president of the company very publicly and very clearly fires you from your job as jellybean counter because of your gross incompetence.
Surely, such a thing could never happen, right?
Unfortunately, if we equate each jellybean with 100,000 jobs in the US economy, this is exactly what happened to Erika McEntarfer, the now-former BLS Commissioner, just last week.
Consider that there are roughly 159 million people currently employed in nonfarm jobs alone in the United States today. The change in employment would be the difference between last month’s jobs number and this month’s jobs number. Here is a table with the preliminary estimates of the total number of jobs compared to the revised estimates for the past three months:
July
June
May
Preliminary
159,539,000
159,724,000
159,577,000
Revised
—
159,466,000
159,452,000
To calculate these numbers, I used the total number of nonfarm employees in April (since that month did not see any revisions this time around), which was 159,433,000, and then added the preliminary estimates for May (+144,000) to arrive at May’s preliminary total. To calculate June’s preliminary total, I added the preliminary estimate of jobs added in June (+133,000) to the preliminary total for May.
For the revised numbers, I again started with April’s total number of nonfarm employees and added the revised job growth figures (+19,000 in May and +14,000 in June). July’s now-preliminary total comes straight from the BLS reporting that was released on Friday, August 1, 2025.
For May, BLS officials overestimated the total number of nonfarm jobs in the US by a whopping 0.07 percent. For June, they overestimated it by a 0.16 percent. For an agency struggling with budget cuts forcing the Bureau to use smaller sample sizes and overall declining response rates to the household and establishment surveys, to be less than one percent off is a phenomenal achievement.
In fact, there’s strong evidence that the BLS has only gotten more accurate over time, not less. Ernie Tedeschi of The Yale Budget Lab and former chief economist for the White House Council of Economic Advisors released this analysis the trends of revisions:
It is certainly true that these revisions come at a particularly inopportune time for the Trump Administration. One week ago, the Administration and members of the New Right were extolling the greatness of the American economy and pointing to jobs numbers and GDP growth rates as positive evidence. The commentariat was questioning economists and economics as a field, saying that it was time to end the tariff skepticism and embrace the wisdom of Trump.
Labeling these figures as “RIGGED [sic]” is a public relations Band-Aid. But if more jobs numbers like these continue to come in, even with a new, Trump-appointed head of the BLS at the wheel, it will be difficult for the tariff proponents to recover.
But if the numbers turn around and are “revised” upward, doubt will pervade the econosphere, as people will lose faith in the accuracy of the BLS and their ability to correctly interpret what is going on in the economy.
The truth is that we need accurate and politically unbiased numbers from the BLS. As has been shown, they have an incredible track record. This reputation must be maintained if we are to continue to understand the economy and make informed decisions.