Reduction of jobs by 911,000 puts spotlight on BLS and Biden’s legacy

The recent release of revised employment statistics — in which 911,000 jobs thought to have been created over the course of a year seemingly vanished — sharpened the debate over how well President Joe Biden steered the economy. 

It also prompted questions about why the federal agency responsible was so far off in its previous estimates. 

The Bureau of Labor Statistics, the government’s official arbiter of employment data, reported Sept. 9 that 911,000 fewer jobs had been created between March 2024 and March 2025 than initial data reflected. The downward revision was bigger^[1] than many economists had predicted and ranks among the largest such revisions in recent decades. 

President Donald Trump’s White House pointed to the BLS announcement as evidence of poor economic stewardship^[2] by Biden, and justification for firing^[3] BLS Commissioner Erika McEntarfer on Aug. 1.

“Biden’s economy was a disaster and the BLS is broken,” the White House said^[4]. “This is exactly why we need new leadership to restore trust and confidence in the BLS’s data on behalf of the financial markets, businesses, policymakers, and families that rely on this data to make major decisions.”

Vice President JD Vance said in an X post^[5], “It’s difficult to overstate how useless BLS data had become. A change was necessary (to) restore confidence.”

But economists told PolitiFact that the lessons of the Sept. 9 revision are more nuanced than that. 

The disappearance of 911,000 jobs is undoubtedly a blow to Biden’s economic legacy, which was already hurt by inflation hitting a 40-year high before returning to levels that were closer to normal.

Still, the BLS process that produced that number, while imperfect, was transparent and has been used by the agency for decades.

“There are always things that can be done to improve the data, but they are incremental and they cost money,” said Dean Baker, co-founder of the liberal Center for Economic and Policy Research. “We can try to reduce the errors with larger surveys, but that will take more money at a time when BLS is seeing its budget sharply reduced.”

What does the revision mean for the Biden administration’s economic record, and for BLS? Let’s review. 

Why does BLS make revisions like this?

With jobs reports, there’s a tradeoff between faster data and more complete data. 

The monthly employment statistics — which employers and economists watch closely when making investment and hiring decisions — are based on payroll data that employers submit to BLS. But some employers submit that information later than others, and so the initial monthly employment figures can change one or two months after they are reported, as more employers send in their data.

A separate revision — the benchmark revision just announced — occurs once a year. A preliminary benchmark revision is made first, followed by a final announcement early the following year. (The preliminary revision for the year ending in March 2024 reported^[6] a loss of 818,000 jobs, and the final number ended up smaller but still large — a loss of 598,000 jobs, a few months later.)

The benchmark revision draws from data that’s considered more finalized and reliable, including information from a quarterly employment study and from unemployment insurance data. 

“There is no political bias at work in the revision,” Michael Strain, the director of economic policy studies at the conservative American Enterprise Institute, said on X^[7]. “This is a standard part of data production that long predates the Trump administration.”

How big was this year’s preliminary benchmark revision?

The preliminary revision of 911,000 jobs this year is big by historical standards.

“This is a large revision, roughly 0.6% of employment,” said Douglas Holtz-Eakin, president of the center-right American Action Forum. “A typical revision is 0.2% to 0.3%.”

The downgrade erased roughly half the jobs in the BLS’ estimate between March 2024 and March 2025.

If the downgrade holds in the final version, the total job creation on Biden’s watch would decline from about 16.1 million to about 15.2 million.

“It really hurts (the Biden administration’s) rhetoric surrounding jobs,” Holtz-Eakin said. “Now instead of getting jobs at the expense of high inflation, we just got the inflation.” 

That said, a loss of 911,000 jobs represents a fraction of all jobs in an economy of 160 million American workers.

There is one silver lining of the downward revision for Biden’s legacy, Baker said. Productivity — the amount of economic activity per worker — at the end of 2024 should end up being 0.7% higher than reported once the job revisions are factored in.

“In the short term, more rapid productivity growth means lower inflation,” Baker said in a blog post^[8]. “Over the long term, it creates a basis for higher wages and living standards.”

Why was the benchmark revision so big this year? 

Economists trace^[9] the particularly large revision this year to several data-collection challenges.

One involves the “birth-death model” that’s used to produce the employer payroll figures. 

The monthly payroll employment data is based on surveys of employers known to BLS. But during times of rapid economic change — either positive or negative — this method is less successful at capturing companies that are “born” or “die” during that period. These figures must be estimated, and those estimates can be off.

Another complication involves workers who are in the United States illegally. Such workers may be counted in the employer payroll survey, because that survey doesn’t ask about immigration status, but they might not be counted in the unemployment insurance data used for the benchmark revision because they may not qualify for unemployment insurance. 

“Immigration slowed sharply in June under Biden, and then further after Trump took office,” Baker said. “Fewer immigrants means fewer workers and jobs.”

Finally, economists have been worried^[10] in recent years about declining response rates by employers who take part in the payroll surveys.

Is the process broken?

Economists told PolitiFact that they distinguish between Trump’s baseless^[11] allegations that BLS numbers are “rigged” for political reasons and more legitimate concerns about the agency’s ability to keep pace with U.S. employment trends.

“While the recent benchmark revisions are significant, they are likely the result of changes in the larger economy, as is widely recognized among people who follow labor-market data, rather than bias in the reporting of the data,” the Council of Professional Associations on Federal Statistics said in a statement^[12] after the preliminary benchmark revision’s release.

One concern is that the monthly dataset’s sample sizes are getting too small. 

“They need more data on a timely basis,” Holtz-Eakin said. But he cautioned that Trump’s nomination of E.J. Antoni, a conservative Heritage Foundation economist, to replace McEntarfer will not solve that problem.

Another concern is the bureau’s dwindling resources under the Trump administration. The agency’s staff count is down by about 20% since February, and a third of its top posts are vacant, former BLS Commissioner Erica Groshen told CNN^[13]. This has led to reductions in some longstanding duties, such as collection of price data for inflation measurements, she said.

Timelier data collection could help improve employment data, but that would require a bigger budget for the agency, said Gary Burtless, an economist at the Brookings Institution. But a larger BLS budget may not be realistic, and states might balk at stricter deadlines for submission of unemployment insurance and payroll statistics, he said.

“Improvements like these are very unlikely to be implemented,” Burtless said.