U.S. nonfarm payroll employment fell -92,000 in February while the unemployment rate rose +0.1 ppt to 4.4%, signaling a clear slowdown in labor market momentum after stronger hiring earlier in the year.

• Nonfarm payrolls declined -92,000 in February versus expectations for +59,000, marking the worst monthly job loss since October 2025 (-140,000) and the second weakest reading since December 2020.

• Private sector employment fell -86,000, the largest decline since December 2020, with job losses occurring across most industries.

• Goods-producing employment declined -25,000 in February, including construction (-11,000) and manufacturing (-12,000), bringing total goods sector job growth over the past three months to -5,000.

• Services employment dropped -61,000 after rising +95,000 in January, with major declines in education & health services (-34,000), leisure & hospitality (-27,000), information (-11,000), and transportation & warehousing (-11,300).

• A strike in physician offices reduced health care employment by -37,000; excluding this effect would raise the education & health services change to +3,000 and adjust total payroll losses to roughly -55,000.

• Government employment declined -6,000 in February, including a -10,000 drop in federal jobs; federal employment is now down -330,000 from its October 2024 peak, a decline of more than -10%.

• Revisions lowered prior payroll estimates by a combined -69,000, with December revised from +48,000 to -17,000 and January reduced -4,000 to +126,000.

• The three-month average payroll gain slowed to +6,000 while the six-month average turned negative at -1,000, indicating a notable loss of hiring momentum.

• The unemployment rate increased to 4.4% (+0.1 ppt MoM), with employment falling -185,000 and unemployment rising +203,000 while the labor force increased only +18,000.

• The labor force participation rate declined -0.1 ppt to 62.0% and the employment-population ratio fell -0.1 ppt to 59.3%, reflecting fewer individuals employed relative to the population.

• Annual population revisions lowered both the labor force and employment levels by about -1.4 million, while leaving the unemployment rate trend largely unchanged.

• Participation among workers aged 55+ declined to 37.3%, the lowest since 2005, while prime-age participation (25–54) remained relatively stable despite the broader demographic adjustments.

• The annual growth in average hourly earnings accelerated slightly from 3.7% YoY in January to 3.8% YoY in February, with goods wage growth up to 4.2% YoY, the highest since March 2025.

U.S. nonfarm payrolls rose +130,000 in January (vs 70,000 expected) while the unemployment rate fell -0.1 ppt to 4.3%, signaling renewed labor market stability at the start of 2026.

• Nonfarm payrolls increased +130,000 in January, the strongest gain since April 2025, lifting the 3-month average to +73,000 after negative prints in November (-8,000) and December (-17,000), indicating a rebound in hiring momentum.

• Private payrolls rose +172,000, offsetting a -42,000 decline in government employment; federal payrolls fell -34,000 and are now down -327,000 (-10.9%) from their October 2024 peak, showing continued public-sector drag.

• Within private hiring, goods-producing jobs increased +36,000, led by construction (+33,000), while manufacturing added +5,000 after prior weakness; services added +136,000, with health care and social assistance contributing +123,500, underscoring concentration in one sector.

• Annual benchmark revisions lowered March 2025 payroll levels by -898,000 SA, cutting 2025 job growth from +584,000 to +181,000, marking the weakest year since 2003 and indicating prior labor strength was overstated.

• The unemployment rate declined to 4.3% (-0.1 ppt MoM; -0.26 ppts since November but +0.26 ppts YoY), as employment rose +528,000 and the labor force increased +387,000; both participation (62.5%) and the employment-population ratio (59.8%) rose +0.1 ppts.

• The number of unemployed fell -141,000, long-term unemployment declined -113,000, and part-time for economic reasons dropped -453,000, pointing to improvement in broader underemployment measures.

• Prime-age (25–54) unemployment held at 3.7% (0.0 ppts MoM), while the 25–34 cohort rose +0.5 ppts to 4.8%, and youth unemployment (16–19) fell -1.9 ppts MoM to 13.6% but remains about +2 ppts YoY, highlighting mixed age dynamics.

• Average hourly earnings rose +0.4% MoM and +3.7% YoY (unchanged YoY), with goods wage growth easing to +4.0% YoY (from +4.2%) and services near +3.65% YoY, maintaining wage growth within the 3.5%–4.0% range.

Establishment Data

US nonfarm payroll employment increased by 50,000 in December, meaning that the year ended with low job growth. The gain was below the estimate of an increase of 60,000 and slightly below the 56,000 increase in November. Revisions of October and November data were less flattering. The October nonfarm employment decline was revised down to -173,000 (from -105,000 in the 1st estimate), and the November gain was revised down to +56,000 (from +64,000 in the 1st estimate). In total, there were -76,000 jobs lost to revisions. Indeed, the negative bias to revisions reversed the trend of a slight recovery in the 3-month average in November to job losses all throughout Q4 2025.

The pace of private sector hiring weakened in December, dropping to 37,000 in December after a 50,000 rise in November. This was below expectations of a 64,000 increase and slightly weaker than the ADP private employment gain of 41,000. Like the headline job growth number, private sector job gains were very affected by revisions. The October increase of 52,000 was revised down to 1,000, and the November increase of 69,000 was revised down to 50,000. This trend reduced the momentum in private hiring (measured by the 3-month average) to 29,000. well below the November print in last month’s report of 75,000. Here are some details in the industry breakdown:

• The goods sector lost -21,000 jobs in December, led by a decline of -11,000 in the construction industry. Every industry within the goods sector saw employment fall. Manufacturing employment dropped by -8,000, the eighth straight monthly decline.
• The service sector added 58,000 jobs in December, continuing to be the main source of job gains in the private sector. There was more variability in the industries, with about half losing jobs (retail trade down the most at -25,000) and about half adding jobs (leisure & hospitality added the most at +47,000). Education & health services maintained its streak of consistent solid gain, adding 41,000 in December.
• The diffusion index, measuring the balance between gaining and losing industries, was weak in December. The index for all private industries was down to 50.8, indicating that the gaining industries barely had an edge. The index for all manufacturing industries was an exceptionally weak 38.2, the lowest since May and the 3rd lowest in over five years.

Public sector employment was back to increasing in December, but not nearly enough to offset the massive October decline. The final month of the year saw government jobs up 13,000, leaving total public sector employment down -155,000 in Q4 2025 alone. This large drop can be attributed to several factors impacting federal government employment, including the government shutdown and the DOGE early retirement program. The October drop of -173,000 is only a one-time shift in government employment, and I expect it to return to increasing in 2026, albeit slowly increasing.

Overall, the Establishment data tells the story of a weak labor market in Q4 2025, but the large decline in government employment in October does obfuscate the trend a bit. Private sector employment reflects a more true picture of where the labor market is going into the new year, a slow deterioration in hiring, but no real capitulation to layoffs. This reflects the descriptions of a “no hire, no fire” employment situation in the US, described by Fed officials, including Chair Powell.

Household Data

The Household data has been the more volatile, uncertain section of the jobs report due to the disruptions from the government shutdown, and now because of the seasonal adjustment refactoring in the December report. The latest data point for the unemployment rate is back below 4.40% with a December reading of 4.38%, the lowest since August 2025. This was down -16 bps from the 4.54% reading in November and below expectations of analysts’ forecast of 4.50%. It is really important to look at these numbers with 2 decimals of precision, because rounding makes the swings look more significant than they are. At that precision, the 1st estimate of the November unemployment rate of 4.6% was revised down to 4.5% due to new seasonal adjustment (when more precisely, the revision was -2 bps), and the December rate fell to 4.4% (when more precisely, the decline was -14 bps).

The takeaway from all this is that the labor market does not look as loose as it was after last month’s report. Indeed, the other segments oft he labor force flows in December all point to this. The total number of unemployed fell by -278,000, and the total number of employed increased by 232,000 (pushing the employment-population ratio up 0.1 ppts to 59.7%). The labor force shrank by -46,000, and the participation rate dropped -0.1 ppts to 62.5%. Additionally, the short-term unemployment population dropped significantly: less than 5 weeks down -251,000 and 5 to 14 weeks down -121,000. Most month-to-month statistics in this report run against the November narrative that labor market slack had increased significantly in November; however, they push back against the magnitude and not necessarily the direction.

Another good example of the reversal in the narrative is the snap back in the prime age unemployment rate in December. After rising to 3.9% from September to November (the second highest in four years), the BLS reported a -0.2 ppts drop back to 3.7% in December. It’s unclear if this reflects some sort of reality in the US labor market or if the volatility is due to disruptions to data collection and statistical quirks. Regardless, it alters the narrative of accelerated labor market loosening that was developed from the November report.

Wage Growth

Despite signs of weakness in the hiring data from the Establishment survey, wage growth surprised to the upside. Average hourly earnings increased 0.3% MoM and 3.8% YoY in December, above the 3.6% YoY increase that was expected. This was partly due to revisions, however, as the November YoY wage growth was revised up from 3.5% YoY in the 1st estimate to 3.6% YoY in the 2nd estimate. Both the goods and services sectors saw wage growth accelerate in December, and YoY growth in average hourly earnings in the goods sector specifically accelerated to a nine-month high of 4.1%.

Establishment Data

The report does provide an update on the Establishment dataset for both October and November, as the data was still able to be collected despite the government shutdown. Total nonfarm payroll employment was shown to have increased by 64,000 in November, after a -105,000 drop in October. On balance, this suggests that -41,000 jobs were lost in the two-month period since September. However, the government shutdown heavily affected government payrolls. While furloughed workers were not counted as unemployed, the BLS counted any contractors or government employees who did not work/receive pay in October as unemployed. This resulted in a -157,000 decline in government payrolls.

Private payroll growth tells a less volatile story. Private payrolls increased in both of the reported months, by 52,000 in October and by 69,000 in November. Notably, the November increase is above the 45,000 that was forecasted by analysts and completely different than the -32,000 reported in ADP’s National Employment Report. The stronger data point means that there was an increase in hiring momentum as the 3-month moving average of private payroll growth increased from 57,000 in September to 75,000 in November. Some notes on the industry breakdown:

• The goods sector added 19,000 jobs in November after losing -9,000 in October. Basically, all of the growth came from an increase of 28,000 in the construction industry in November. The manufacturing sector continues to see employment decline: down by -9,000 in October and by -5,000 in November.
• The services sector added 50,000 jobs in November and 61,000 jobs in October. Like in the goods sector, most of the growth across both months came from one segment, education and health services, which added 65,000 in November and 59,000 in October. That means that without this segment, service sector employment would’ve contracted in the last 2 months.
• Industries that might be positively impacted by the holiday period appeared to be in a slump. The wholesale and retail trade together only added 3,900 jobs in October and November, and the transportation & warehousing industry reported payrolls down by -18,200 in that period.

As mentioned before, government payrolls have become a bit of a confounding variable in this jobs report due to the volatility caused by the government shutdown. Government payrolls dropped -162,000 across both months, a very unusual result, but the BLS does explain some of it. Not only was the decline affected by the government shutdown, but there was also a decline in employment as a result of the DOGE deferred resignation offer for federal employees. In summary, the DOGE early retirements, disengagement of federal contractors, and temporary layoff of some government workers explain the large October drop. However, the BLS is not able to “precisely quantify” how these factors affected the data. Thus, the only real value in the Establishment data in the November report is in the private payrolls.

Household Data

The November report does not include October data for the Household survey since it was not collected. Thus, we only have November data points to look at and 2-month deltas. The headline unemployment rate ticked up 12 bps to 4.56% from September to November, resulting in a rounded print of 4.6% (up from a rounded 4.4%). The rounded change of 0.2 ppts slightly overstates the increase in the unemployment rate, which is actually just over half that. Nevertheless, there was a material increase in joblessness after September, such that the unemployment rate is now at the highest point since September 2021.

The increase in the unemployment rate came from an increase in participation that led to a larger portion of that increase in the labor force becoming unemployed versus employed. Specifically, the labor force increased by 323,000 from September to November (participation rate up 0.1 ppts to 62.5%). The employed population increased by just 96,000 in that period (employment-population ratio down -0.1 ppt to 59.6%) while the unemployed population increased by 228,000 in the same period.

The BLS does note that the November Household survey should be interpreted with added caution, given several shutdown-related disruptions.

• October data were not collected and were not backfilled, meaning November estimates rely on a modified weighting process that shifts previously collected data forward and effectively forces analysis into two-month changes rather than standard month-to-month comparisons.
• Data collection for November began late and was extended around the Thanksgiving holiday, resulting in a lower-than-usual response rate of 64 percent and slightly higher standard errors across key series. While seasonal adjustment methods were unchanged and the reference week was otherwise typical, these factors increase statistical noise and reduce precision.
• Federal government workers were counted as employed because the government reopened before the end of the reference week, and the BLS noted that the overall impact of the shutdown on Household survey estimates cannot be precisely quantified.

Despite these caveats, the Household data are a clear indicator of weakening in the labor market, and much more convincing than the Establishment data. The increase in joblessness was seen across several age groups, including the key prime working age cohort (25 to 54 years), where the unemployment rate increased 0.2 ppts to 3.9% over the two-month period, the highest it has been since November 2024 and, before that, since November 2021. Additionally, the part-time work for economic reasons due to slack work/business conditions jumped 321,000 in the two-month period, or an average of 160,500 in each month, which would be the highest monthly increases since February. It’s unlikely this part-time series was impacted by the government shutdown since the “only could find part-time work” indicator captured a roughly 600,000 increase that was likely caused by the federal work disruption.

Establishment Survey

US nonfarm payroll employment increased by 119,000 in September, a surprise burst of strength exhibited by the labor market at the end of Q3. Not only was this increase above consensus expectations of a 50,000 increase, it was the largest increase since April and the third largest increase so far this year (just below March gain of 120,000). Most significantly, the strong growth in September reverses a string of weak readings that began in May and lasted through the summer, and resulted in job growth averaging around 18,500 over the course of those four months. Revisions in today’s report made those numbers look even worse as the July total was taken down -7,000 to 72,000, and the August total was revised down -26,000, such that employment fell -2,000 during the month.

The headline number was supported by the strongest private payroll growth since April. The 97,000 jobs added by private sector firms were ahead of expectations of a 62,000 increase and substantially above the September ADP private sector employment change estimate of -29,000. As usual, the service sector provided most of the job growth, adding 87,000, while the goods sector only added 10,000. Two services industries specifically accounted for almost all of the increase in nonfarm payrolls: education and health services (+59,000) and leisure and hospitality (+47,000). Job growth in other industries was a wash. Notably, with the Trump admin’s focus on tariffs in mind, manufacturing employment fell for the 5th straight month since the reciprocal tariffs were put in place in April (down -58,000 cumulative since then).

Government job growth made a small positive contribution at +22,000. This included federal employment down -3,000, offset by 25,000 jobs added in state & local governments. This, of course, doesn’t include the job losses that may have come about as a result of the reduction in force orders during the government shutdown. Jobless claims data suggest that the number was not insignificant. Initial claims by federal employees totaled 5,719 in the week of November 8th, well above 615 in the same week a year ago.

Taken as given, the data from the Establishment survey looked pretty solid. Hiring appears to have bounced back after a weak summer, and the private sector accounted for a majority of the increase. The breadth of job gains across industries was not great, but there didn’t appear to be evidence of widespread layoffs. The one caveat is that these numbers are likely to be revised, and revisions have not been friendly.

Household Survey

If you were optimistic reading through the Establishment survey data, this is the part of the report where things get ugly. The headline unemployment rate saw a significant increase in September, rising 13 bps to 4.44%, above expectations of it remaining at 4.3% and the highest since October 2021. This increase wasn’t a matter of rounding. The print came within 1.1 bps of rounding up to 4.5%. It sounds nitpicky to comment at this level of detail, but it is important to note these things since a majority of people view these data points as a single data point. The takeaway is that there was a solid, measurable increase in unemployment in September, confirming a clear trend that has been established this year.

The increase in the unemployment rate came from an increase in the total number of unemployed (+219,000) that resulted from a sizeable increase in the civilian labor force (+470,000). While there was an increase in the employed population (+251,000) and the employment rate (+0.1 ppts to 59.7%), the raw increase in the unemployed was just too high relative to the increase in the number of employed. Interestingly, this does contradict some Fed officials’ views on the labor market that suggest labor supply and labor demand are both going down, leading to equilibrium rather than slack.

The age breakdown shows that prime-age and older workers drove the entire increase in September’s unemployment rate. The youngest cohort, 16 to 19 years, actually saw a notable decline in their jobless rate to 13.2%, and workers aged 20 to 24 held steady at 9.2%. With younger groups either improving or unchanged, the pressure came from older segments. The prime age unemployment rate rose 0.1 ppts to 3.5%, the 55 and over rate jumped 0.4 ppts to 3.3%, and early prime age workers posted the sharpest move with a 0.4 ppt increase to 4.8%. This pattern points to a weaker labor market because job losses are increasingly concentrated among core working-age adults rather than temporary or seasonal young workers, signaling more persistent underlying softness rather than noise in the data.

The Household survey data make it clear that unemployment moved meaningfully higher in September. On top of that, the increase in the labor force and in the participation rate push back against the notion that the labor market is in a sort of equilibrium state where labor supply and labor demand are both falling in tandem, such that there was nothing more than a slow cooling in the labor market.