EL7 Research · Macroeconomics
29 September 2026 · Data through noon on 30 September · Report due Friday, 2 October, at 12:30 GMT
The September US jobs report is due Friday, with consensus expecting 90 thousand jobs. We broke down August’s 162 thousand figure component by component, built our forecast for each, then calculated the two rounding thresholds that will determine the report’s reading: unemployment stands at 4.14% before rounding, while hourly earnings would lift annual growth to 3.2% at $37.86. Using EL7’s price database, we measured how markets moved around 33 previous reports and what the figure means for the Federal Reserve’s 28 October decision.
Our forecast: 70 thousand new jobs in September, slightly below consensus estimates (Consensus Estimates), which reflect the prevailing view among analysts. The September US jobs report is due Friday, 2 October 2026, at 12:30 GMT. Consensus estimates in EL7’s calendar have risen to 90 thousand, matching other surveys, with a range of 35 thousand to 180 thousand. We build our forecast component by component and assign it a range of 15 thousand to 125 thousand — EL7 estimates.
We retained our payroll forecast after three indicators released since publication and revised our unemployment rate (Unemployment Rate) forecast, the share of the labour force without work. The private sector added 90 thousand jobs in the September ADP report, above the 70 thousand forecast. In the August Job Openings survey (Job Openings (JOLTS), hires (Hires), the number of people who started jobs during the month, rose to 5.192 million. Layoffs and discharges (Layoffs and Discharges), meaning employer-initiated terminations, fell to 1.641 million. Job openings, available positions not yet filled, declined to 7.079 million. By contrast, the labour differential (Labor Differential), the gap between the shares of respondents saying jobs are plentiful and hard to get, fell in The Conference Board Consumer Confidence Index to +1.7 points, its lowest since February 2021. In other words, company-based employment indicators did not weaken, while households’ assessment of the labour market deteriorated clearly. That assessment typically appears in the household survey (Household Survey), which polls individuals and is used to calculate the unemployment rate.
We do not expect August’s distribution of job gains to be repeated. The economy added 162 thousand jobs in August, but restaurants and bars alone added 59.2 thousand, about four times their monthly average, while local-government education employment increased by 41.9 thousand after falling in July. Excluding those two categories and healthcare leaves the cyclical private sector: private jobs most exposed to fluctuations in economic activity. Its additions averaged only 34 thousand jobs over three months.
The unemployment rate (Unemployment Rate) is the report’s most sensitive figure. The published rate is 4.1%, but it is 4.14% before rounding. If the labour force remained unchanged, an increase of roughly 15 thousand unemployed people would be enough for the rate to be published at 4.2%. With the share saying jobs are “hard to get” rising to 21.9% in September, the highest since January 2021, we now assign a 50% probability to the published rate reaching 4.2% or higher and a 42% probability to it remaining at 4.1% — EL7 estimates.
Wage growth is at a similar rounding threshold. Average hourly earnings (Average Hourly Earnings), the average amount a private-sector worker receives for each hour worked, reached $37.75 in August. If they reach $37.86 or more in September, annual growth will be published at 3.2%, not 3.1%. We forecast a monthly increase of 0.2% to $37.84, implying annual growth of 3.1%.
Friday’s report comes 26 days before the Federal Reserve (Federal Reserve) meeting. The US central bank’s CME FedWatch Tool (CME (CME FedWatch Tool) derives decision probabilities from federal funds futures (Fed Funds Futures), contracts that reflect expectations for the federal funds rate. According to the tool, the probability of another rate increase on 28 October was 34.9% at noon on 30 September, down from 70.9% a week earlier. The decline came in two stages: first after John Williams, president of the Federal Reserve Bank of New York, said there was “no need to rush”; and second after core inflation (Core Inflation), which excludes food and energy, came in at an annual 3.0% in the Personal Consumption Expenditures measure (Personal Consumption Expenditures (PCE)), which tracks changes in the prices of goods and services consumed by individuals. The same reading is corroborated by SOFR futures (SOFR Futures) and the decline in the two-year Treasury yield, while the Polymarket prediction market gives a similar figure of about 34%. A strong payroll reading with unemployment at 4.1% could raise the probability of another rate increase, while a weak reading with unemployment at 4.2% would reduce it further. But our measurement of 33 reports since 2024 shows that the initial market reaction reverses in roughly half of cases during the first hour.
A free account gives you 4 full research papers every month. Or subscribe to EL7 Pro for full access to more than 300 papers.
$700 per year · or $99 per month