Another blow to US employment: the annual review through March 2025 eliminates 911,000 jobs in one fell swoop.

As was the case last year and as expected, the US created significantly fewer jobs in the 12 months prior to March than initially announced. The preliminary revision published this Tuesday by the Bureau of Labor Statistics (BLS) reflects that from April 2024 to March 2025 (inclusive) 911,000 fewer non-farm jobs were created than previously reported . Although economists had expected a "bulk" figure, and despite the fact that this revision will not be incorporated into official data until next year (February 2026), the "hole" increases pressure on the Federal Reserve to gallop down interest rates. If the weak August employment report , published on Friday, opened the debate about a "jumbo" 50 basis point cut by the Fed next week, this revision could exacerbate it.

After the uproar over significant downward revisions to nonfarm payrolls in May and June , which cost the head of the BLS her job after Trump abruptly fired her , a figure of between 500,000 and 800,000 had been expected for this statistical adjustment, which is made each year by cross-referencing data from the monthly business survey with data from the Quarterly Census of Employment and Wages (QCEW). The QCEW data is based on state unemployment insurance tax records that nearly all employers are required to file. Before Tuesday's report , government payroll data showed that employers added nearly 1.8 million jobs total in the year through March on a non-seasonally adjusted basis, or an average of 149,000 per month.

Fed Governor Christopher Waller, the central bank's most vocal advocate for rate cuts, hinted at a figure close to 720,000 fewer jobs in his speech just over a week ago, which implied a reduction of 60,000 positions per month. The data shows that he fell short ( more than 70,000 are "disappearing" on average each month ), which will fuel debate within the central bank. "This would represent a significant shift in the labor market narrative and could raise questions about why the Fed shouldn't cut 50 basis points this month," ING analysts note.

"Most of the downward revision appeared to occur during the last half of the period from October 2024 to March 2025. If this trend continues in more recent months, it increases the risk that the worrying 500,000 decline in the employment indicator of the Household Survey (the other leg of the official employment survey alongside the Business Survey) over the past seven months provides a more accurate guide to labor market conditions," Capital Economics assesses.

The breakdown shows widespread revisions, with the largest reductions in several service sectors, specifically leisure and hospitality (15,000 fewer jobs per month), professional and business services (13,000), and retail trade (10,000). "Given that services are the last bastion of job growth, this does not bode well for the overall health of the labor market," the analyst firm concludes.

As SEB points out, the data in this revision is "outdated," as it only covers the period up to March. "Recent data—including downward revisions to payrolls in May and June and the fact that employment has been driven solely by non-cyclical hiring, particularly in the healthcare sector—already indicate a clear slowdown in job growth, following an unexpected acceleration at the end of last year and again during the spring," they clarify.

This does not negate the impact on interest rate expectations . In 2024, the final revision resulted in the "elimination" of 600,000 jobs. "Such sweeping downward revisions have rarely been observed outside of recessions , which reinforces the focus on these data," the Swedish bank explains.

One reason for these variations is the BLS's assumptions about the net employment impact of new business creation and the loss of business closures ( the birth-death model ), according to Nordic Bank's research department. "These models can be particularly misleading around economic turning points," they argue.

"The US is offering benchmark revisions to employment data: with survey response rates falling, real-time data is becoming less grounded in reality. Some notable negative revisions are expected. The lower number of people employed in the past shouldn't change the Federal Reserve's outlook (the overall economic outlook, including inflation, is unchanged). However, if the revisions indicate worsening employment in the immediate past, that negative momentum should affect Fed policy," noted Paul Donovan, chief economist at UBS, before the figure was released.

The BLS adjustment indicates that the labor market slowdown in recent months has followed a prolonged period of more moderate job growth, which could set the stage for a series of interest rate cuts starting next week. Federal Reserve Chairman Jerome Powell recently acknowledged that risks to the labor market have increased , and two of his colleagues opted to reduce borrowing costs in July.