Setting the Record Straight

Trump fired the head of the BLS and told three whoppers in the process.

Photos courtesy of ABC News

On Friday, after Trump was shown the revised job numbers revealing the economy was not nearly as strong as he had been touting, Trump promptly fired the messenger. In this case it was Erika McEntarfer, head of the Bureau of Labor Statistics (BLS), a career civil servant who had won broad, bipartisan support in her Senate confirmation.

Trump’s move was not only petty and vindictive, it was a message to future holders of that and other positions: Fudge the numbers my way, or you’ll suffer the same fate.

But Trump isn’t just interested in controlling the messaging and numbers going forward. He’s also keenly interested in rewriting the past, all in a bizarre yet dangerous effort to justify his removal of an official for, essentially, doing her job.

Trump told three bald-faced lies in service of that effort, most of which the media has failed to correct. So let’s set the record straight, because the way things actually work could trip Trump up badly going forward.

No, the job numbers aren’t “rigged”

First, Trump asserted in a post on his increasingly ironically named platform, Truth Social, that the job figures were “rigged” to make him look bad.

He repeated this lie on Sunday.

The numbers that come in on jobs involve whole teams of data collectors and surveys with over 100,000 employer respondents each month. Specifically, according to its website, BLS surveys some 121,000 businesses and government agencies to get a reported headcount as of the 12th of each month. This survey results in the initial non-farm payroll numbers we often hear about.

So why does that preliminary number change? BLS applies a seasonal adjustment, a birth‑death model to account for new business formation/closures not yet in the sample, and, importantly, later revisions as more reports arrive. Many survey respondents aren’t turning in their surveys in a timely manner, and when they do, numbers can shift, often dramatically. The tardiness of replies is especially prevalent in this post-pandemic environment.

Given how dispersed this work is, in order to “rig” these numbers the head of the BLS would have had to conspire with data collectors to alter the numbers as they came in. Such a move would be easily detectable, and very likely someone would whistleblow any effort to do that. So let’s be clear about what he’s really after. Trump blasting the numbers as “rigged,” with no evidence or understanding of how the numbers get calculated, is just another way to sow doubt in the minds of the uninformed. And it advances his preferred narrative that the economy is strong when it is actually teetering precariously.

By the way, it won’t be easy for McEntarfer’s replacement to fudge numbers either, at least not without massively changing the way the data gets collected and tabulated. We should watch for any effort to inject subjectivity into what is essentially a dry, data-driven process.

No, the numbers were not revised “massively downward” AFTER the election

A second lie Trump told to justify his firing of McEntarfer was that she (and only she) had somehow revised the 2024 job numbers in such a way as to benefit Kamala Harris. Specifically, Trump claimed that the 2024 job numbers were “massively revised DOWNWARD” by 818,000 after the election and that this was “A TOTAL SCAM.”

But even a basic review of the timeline shows this claim is dead wrong. In fact, the big downward revision occurred in August of 2024, months before the election and at a very bad time for the Democrats. So no, they did not happen after the election as Trump falsely claimed.

As CNN reported at the time last August:

The Bureau of Labor Statistics’ preliminary annual benchmark review of employment data suggests that there were 818,000 fewer jobs in March of this year than were initially reported.

It noted that such a possible revision happens each year:

Every year, the BLS conducts a revision to the data from its monthly survey of businesses’ payrolls, then benchmarks the March employment level to those measured by the Quarterly Census of Employment and Wages program.

And the news was not great for the Democrats going into the fall election:

The preliminary data marks the largest downward revision since 2009 and shows that the labor market wasn’t quite as red hot as initially thought. However, job growth was still historically strong.

If McEntarfer were trying to help the Democrats, she wouldn’t have issued her report the way she did. As is, it helped destroy the narrative that there was great strong job growth under Biden just months before the election.

Despite the clear timeline, which has been pointed out in the media, Trump doubled down on his false claims, telling reporters that McEntarfer had rigged the announcements. As The Guardian reported:

“Before the election,” Trump recalled, wrongly, “she gave out numbers that were so good for the Democrats, it was like unbelievable.”

“And then right after the election, she corrected those numbers with, I think, almost 900,000 correction,” he said, referring incorrectly to the revision that had taken place in August and had been a boon to his campaign.

No, Trump didn’t always think the numbers were wrong

To hear Trump and his people tell it today, he has always believed the BLS numbers were wrong, has always hated them, and had always planned to fire McEntarfer over them. He’s trying to establish this narrative so it doesn’t look like what really happened, i.e., that he saw bad job numbers, got angry about them, then punished the person who reported them.

In fact, when the August 2024 revisions came out, Trump had sought to capitalize on them, claiming they proved Biden and Harris had been padding the numbers all along—and he credited the BLS for uncovering the truth.

Trump also loved the very low job numbers reported in early November, just days before the election, showing only 12,000 jobs created. His campaign even called the report “a catastrophe” that “definitively reveals how badly Kamala Harris broke our economy.”

Nor did Trump “hate” the numbers the BLS numbers out earlier this year when they showed stronger job growth than was actually happening. On the contrary, when the “good” numbers came out this summer, Trump positively crowed about them. Meet the Press’s Kristen Welker noted that Trump was “happy to accept” McEntarfer’s numbers when they favored him, and the program put together a montage of the times Trump had amplified these numbers to his political advantage:

I do not envy whatever fool agrees to serve as the next head of the BLS, which at this point ought to just remove the “L” and just go by “BS.” Either they will report numbers to Trump’s liking, or the new head will, well, lose theirs just as McEntarfer did.

Stop the count!

One final point. There’s a pattern to this behavior that we should recognize. When Trump liked the preliminary numbers he was seeing on Election Night in 2020, he claimed an early victory, not wanting the count to change as the night went on and big city ballot dumps and mail-in votes were tabulated. He then used the “change” in the numbers to allege that the system was “rigged” because, in his world view, nothing could explain such a big shift.

The explanation for the shift was as basic as today’s job numbers shift. The change in vote tabulations on Election Night and the days to follow was fully expected, even by the Trump campaign, as the “red mirage” of victory faded the more votes came in. Simply put, elections officials weren’t done counting and tabulating the data. Good, accurate data takes time.

Yet time and again, we see Trump exploiting that gap to his advantage when it suits him. There may come a time, and it may be soon, when government statisticians, election officials and other officials decide it’s advisable to wait and not release preliminary results or information, given the chance that the initial reports will be weaponized in ways unimaginable before MAGA Trumpism.

That is the world we unfortunately now live in. To rebuild the citizenry’s trust in our government, we may need to eliminate all areas of perceived subjectivity or potential mischief, even where there isn’t and has never been evidence of “rigging” or other foul play. This is especially true where the initial results may vary considerably from the final ones, and where the unfocused and unlearned are apt to cry conspiracy because they simply do not and cannot understand basic math or the limitations of data collection.