Every quarter, the Bureau of Labor Statistics publishes a snapshot of how many jobs American businesses create and destroy. The most recent release, for the fourth quarter of 2025, shows a labor market that is constantly churning beneath a calm headline number, and it carries a few caveats that are easy to miss.
What the Business Employment Dynamics Release Actually Measures
The Bureau of Labor Statistics describes Business Employment Dynamics as a set of statistics generated from the Quarterly Census of Employment and Wages program, with quarterly gross job gains and losses tracked from 1992 onward. Instead of reporting only the net change in employment, the series splits it into its moving parts: jobs added by new and growing establishments, and jobs removed by shrinking and closing ones.
The unit is the establishment, meaning a single physical location of a business, not a company. A chain with 40 stores counts as 40 establishments. That matters when reading anything labeled a “business birth”: a new branch of an existing firm can count, and so can a brand-new independent shop.
The release published on July 29, 2026 covers the three months ending in December 2025. All of its main tables are seasonally adjusted, according to the Bureau of Labor Statistics.
The Fourth-Quarter 2025 Numbers
According to the Bureau of Labor Statistics, private-sector gross job gains in the fourth quarter of 2025 were 7,838,000, equal to 5.9% of employment. Gross job losses were 7,245,000, or 5.5%. The difference is a net gain of 593,000 jobs.
| Component | Jobs | Share of employment |
|---|---|---|
| Gross job gains | 7,838,000 | 5.9% |
| From expanding establishments | 6,197,000 | 4.7% |
| From opening establishments | 1,641,000 | 1.2% |
| Gross job losses | 7,245,000 | 5.5% |
| From contracting establishments | 5,812,000 | 4.4% |
| From closing establishments | 1,433,000 | 1.1% |
Two things stand out. First, the gross flows are enormous compared with the net change: roughly 7.8 million jobs were gained and 7.2 million lost in a single quarter, to produce a net figure of about 0.6 million. Second, most of the churn comes from existing establishments growing or shrinking, not from openings and closings. Expanding and contracting sites account for about 12.0 million of the 15.1 million combined gross gains and losses. (Illustrative arithmetic: 6,197,000 plus 5,812,000 is 12,009,000; 7,838,000 plus 7,245,000 is 15,083,000.)
How Much Did the Picture Change From the Prior Quarter?
The previous release, covering the third quarter of 2025, looked quite different. According to the Bureau of Labor Statistics, gross job gains were 7,475,000 (5.6%) and gross job losses were 7,634,000 (5.7%), for a net change of negative 159,000.
Moving from one quarter to the next, gains rose by 363,000 and losses fell by 389,000, consistent with the release’s reported changes. That swing of roughly 750,000 jobs in the net direction, from a small loss to a sizable gain, is a reminder of how much a single quarter can move. Net figures of this kind are noisy, and one quarter rarely settles an argument about whether the economy is strengthening.
For a broader view of how forecasters weigh conflicting signals, see why World Bank and IMF growth forecasts differ.
How Many New Establishments Opened?
The Bureau of Labor Statistics reports that there were 338,000 establishment births in the fourth quarter of 2025, a 3.6% rate, and that these accounted for 1,006,000 jobs, a 0.8% rate. In the third quarter, there were 323,000 births accounting for 967,000 jobs.
A simple division gives a sense of scale. Dividing 1,006,000 jobs by 338,000 births gives about 3.0 jobs per new establishment. The third quarter works out to about the same (967,000 divided by 323,000 is roughly 3.0). That is an average of what these sites employ in the period they appear, and it confirms that the typical new establishment is very small.
A birth, in the Bureau’s technical definition, is an establishment that appears in its longitudinal database for the first time with positive employment in the third month of a quarter. That is a statistical definition, not a legal one, and it does not tell you whether the owner was a first-time founder.
Why Are Closings Reported Later?
Deaths are the other half of the story, and they arrive late. The Bureau states that there is always a lag of three quarters for the publication of death statistics. In the latest release, death figures run only through the first quarter of 2025, when 809,000 jobs were lost at 284,000 establishments, a 3.0% rate. Figures for the second through fourth quarters of 2025 were listed as not available.
The previous release carried deaths through the fourth quarter of 2024: 306,000 establishments and 939,000 jobs lost. According to that release, the jobs-lost figure was corrected in a reissue on June 30, 2026.
This lag creates a common trap. A commentator who sets the latest births (fourth quarter of 2025) against the latest deaths (first quarter of 2025) is comparing different periods. The honest comparison matches the same quarter on both sides, which for now means waiting for the data.
Deaths are also defined narrowly. The Bureau describes them as a subset of closings that does not include temporary shutdowns of seasonal businesses. Because deaths are a subset of closings, closings are always at least as numerous.
Does a Closing Business Mean a Failed Business?
Not necessarily. An establishment can disappear from the data because the owner retired, sold the operation, merged it into another site, or moved it. The statistics record that an establishment’s employment went from positive to zero, not why.
The release also does not report survival rates, so claims such as “most new businesses fail within five years” cannot be tested against this particular release. Readers who see such claims should ask which dataset produced them and over which cohort of businesses.
The Bureau adds a useful perspective of its own: births and deaths, and the jobs they create and destroy, are small in relation to overall gross job gains and losses. This matches the table above. Openings and closings together account for about 3.1 million of the 15.1 million jobs gained or lost, roughly one-fifth. (Illustrative arithmetic: 1,641,000 plus 1,433,000 is 3,074,000, which is about 20% of 15,083,000.)
A Worked Example: Reading One Quarter Without Overreacting
Suppose a local news story says “new business formation soared.” A careful reader can test it in four steps using the figures above:
- Check the unit. Are these establishments or companies? A new branch is a birth.
- Check the period. Births for the fourth quarter of 2025 cannot be netted against deaths for that quarter yet.
- Check the adjustment. The tables discussed here are seasonally adjusted, so comparing them with unadjusted local counts can mislead.
- Check the size. At about 3.0 jobs per birth, a surge of new sites adds relatively few jobs next to the 5.8 million lost at contracting establishments in the same quarter.
The same discipline applies to the question of what owners do when costs climb, which is covered in what small businesses are changing when every cost matters.
What the Data Cannot Tell You
The series is useful but limited. It does not say how profitable new establishments are, whether they were funded by loans or savings, or how long they will operate. It also counts jobs at establishments, not owners’ personal circumstances, so it cannot say who started a business or why.
It also says little about where work happens. Shifts in office use and remote arrangements affect which establishments exist and where, a theme explored in the workplace is still changing.
Used properly, the data supports a narrower, sturdier conclusion: the American job market is a high-turnover system in which a few million jobs appear and vanish every quarter, new establishments are small, and the full closing picture always arrives with a delay.
Reading Business Churn With a Steadier Eye
The fourth-quarter 2025 release shows a net gain of 593,000 jobs built from roughly 7.8 million gains and 7.2 million losses, with 338,000 new establishments averaging about three jobs each. The most dependable habits are to match periods, respect the lag in death data, and remember that most churn comes from existing sites growing and shrinking.
This article is for general information only and is not financial or investment advice.
Sources: Bureau of Labor Statistics
