When the jobs report is released, US markets move in seconds. The number of employed people decreased by 92,000 in February, and the unemployment rate rose to 4.4%, with the previous month’s figure revised downward by 69,000.
Combined, these represent 161,000 fewer jobs than numbers indicated at the beginning of the year.
However, the numbers that traders first react to often do not last long. That’s because an even bigger fix may be coming in the months ahead.
The Bureau of Labor Statistics has already revised down U.S. job growth for the year to March 2025 by 862,000 jobs, raising the possibility that the market and the Federal Reserve are reacting to a job market that looks more robust in the headlines than in the final numbers.
The number of trades in the market is not the final number
This is the real story behind monthly salary announcements. Investors treat the jobs report as one of the most important macro outcomes, and for good reason.
The moment the jobs report is released, Treasury yields change, stock index futures rise, the dollar fluctuates, and expectations for a Fed rate cut or postponement are rewritten within minutes.
However, the numbers that drive the initial response are only estimates. It is based on a survey, revised as more employers respond, and then benchmarked against a broader set of pay records.
This means that labor markets, where traders set prices in real time, are often drafts. Later edits may be small, but they may change the overall picture.
February was weak even before the reset
The February report itself was weak. The BLS announced that total nonfarm payrolls decreased by 92,000 jobs for the month, and the unemployment rate rose to 4.4%. Due in part to the strike action, 28,000 jobs were lost in the healthcare industry, and 37,000 jobs were lost in doctors’ offices alone. The information will result in the loss of 11,000 jobs.
Federal government employment has declined by 10,000 people, and is now down 330,000 from its peak in October 2024. Employment in transportation and warehousing fell by 11,000, while employment in couriers and messengers fell by 17,000.
The report still showed wage growth. Average hourly wages increased by 0.4% in February, and by 3.8% compared to the same month last year.
This is important because it keeps part of the Fed’s inflation problem alive even as employment cools. Wage pressures could emerge even as the labor market weakens, particularly if employment growth slows from the levels that have long supported consumer spending.
However, last month’s revisions weakened the report significantly.
The figure for December was revised from an increase of 48,000 to a decrease of 17,000, and that for January was revised from an increase of 130,000 to a decrease of 126,000.
Together, these changes resulted in a reduction of 69,000 jobs from the previous situation.
Investors are always trying to gauge direction, but the downward revisions indicate the labor market was already losing momentum before the latest report was released.
862,000 employment changes change the situation completely
Then a larger reset occurs. In its annual benchmarking process, the BLS reduced total nonfarm payrolls by 862,000 from March 2025 levels on a non-seasonally adjusted basis. On a seasonally adjusted basis, the revised March 2025 value was 898,000 fewer.
This kind of technical distinction is only important to economists. But the broader conclusion is much simpler. The real-time labor market appeared to be significantly stronger than when the BLS compared survey estimates to more complete employment records.
This large number is not a simple statistical arrangement. This is a reminder that one of the world’s most market-sensitive data releases is not a direct tally of all U.S. jobs. The first number is a high quality quote made with speed in mind. The latter benchmark was built for completeness.
But when the gap between the two becomes this wide, a macro story begins to take shape.
The benchmark changes also change how investors should think about the past year. The real-time appearance of resilience in the labor market helped support the argument that the economy could survive under capped interest rates.
The outlook is increasingly precarious as the labor market turns out to have created far fewer jobs. The data has completely changed the balance of the debate.
Why does the data change so much?
Monthly payroll figures come from the Current Employment Statistics Survey, which samples employers rather than counting all payrolls in the country. This is very large and very useful, but it’s still just a sample.
Monthly revisions occur as additional employer reports arrive after the initial release and seasonal factors are recalculated.
This annual benchmark is primarily based on unemployment insurance tax records and goes further by matching the findings with the Quarterly Employment and Wage Statistics Survey, which covers most payrolls.
That creates inevitable tension in the market. Traders trade quotes because they need numbers quickly. The Fed must leverage the same real-time information, knowing that future revisions may change its shape.
There is no real alternative solution. Some of the biggest market movements each month are based on numbers that may appear to change meaningfully once the data is more complete.
This is why salary revisions are not some vague technical issue. They influence the stories investors tell themselves about growth, inflation, and interest rates. If the labor market looked stronger in the first edition than in the benchmark data, yields, risk sentiment, and interest rate expectations may all have been set against the backdrop of a weaker economy than appears.
Nevertheless, the initial salary amount is still important because it is timely, and timeliness is valuable. But benchmarks exist because the first number is not the final number, and speed and completeness are not the same thing.
The decline in the number of employees in February is important, the unemployment rate rises to 4.4%, and the downward revision to the previous month is important. The indicative 862,000 job cuts may be most important, as the report said the labor market, which shaped much of the macro discussion last year, looked stronger in the headline data than in the aggregate.
In the market, the first number is traded. In labor data, that doesn’t always hold up.
(Tag translation) Bitcoin
