U.S. unemployment claims dip to lowest since mid-July, layoffs remain historically low

U.S. initial unemployment claims fell to a seasonally adjusted 197,000 in the week ending Sept. 26, 2026, down 1,000 from the prior week's revised 198,000 and below the 200,000 consensus forecast in a Reuters poll. The Labor Department's latest revised history makes this the lowest reading since July 18, 2026, when claims stood at 189,000. The four-week moving average fell to 200,000, while continuing claims dropped to 1.701 million in the week ending Sept. 19, their lowest level since April 2023.

August job openings slipped to 7.1 million and layoffs remained at about 1.6 million, while August payrolls rose 162,000 and unemployment held at 4.1 percent. Wage growth, at 3.1 percent from a year earlier, does not by itself suggest a renewed wage-price spiral. Federal Reserve officials nevertheless view labor-market risks as having receded while inflation remains too high, a combination that helped prompt the Fed's Sept. 16 quarter-point rate increase.

The Labor Department reported on Oct. 1 that seasonally adjusted initial claims declined 1,000 to 197,000 for the week ending Sept. 26. The previous week's level was revised upward by 1,000 to 198,000. The initial-claims four-week moving average fell 2,500 to 200,000, with the previous average revised up to 202,500. On an unadjusted basis, initial claims totaled 156,738, down 7,979, or 4.8 percent, from the preceding week; the comparable unadjusted figure a year earlier was 179,162.

Continuing claims, which lag initial claims by a week in the advance release, decreased 11,000 to 1.701 million for the week ending Sept. 19. The prior week's continuing-claims count was revised down by 7,000 to 1.712 million. Their four-week average fell 18,500 to 1,723,750, while the insured unemployment rate held at 1.1 percent. FRED confirms the 1.701 million figure and notes that continuing claims measure people who have already filed an initial claim and then claimed benefits for another week of unemployment.

Reuters reported that 1.701 million was the lowest continuing-claims level since April 2023. The same report said economists had expected 200,000 new claims, making the 197,000 result modestly stronger than forecast.

Because unemployment claims are weekly, the Labor Department does not publish a standard "month-over-month" percentage comparable with monthly indicators such as payrolls. One defensible comparison is the average of the weekly readings: September's four initial-claims observations averaged 200,000, compared with 206,000 across August's five observations, a decline of 6,000, or about 2.9 percent. A complete September monthly average for continuing claims is unspecified, because the latest continuing-claims observation covers only through Sept. 19.

A four-week-apart comparison gives a similar signal. Initial claims fell from 207,000 on Aug. 29 to 197,000 on Sept. 26, a decline of 10,000, or 4.8 percent. Continuing claims fell from 1.775 million on Aug. 22 to 1.701 million on Sept. 19, down 74,000, or 4.2 percent. The initial-claims four-week average declined from 207,500 on Aug. 29 to 200,000 on Sept. 26, or 3.6 percent.

The year-over-year comparisons are more pronounced. Seasonally adjusted initial claims are 28,000, or 12.4 percent, below the comparable 225,000 reading a year earlier. Continuing claims are 220,000, or 11.5 percent, below their comparable 1.921 million year-earlier level, and the initial-claims four-week average is 34,000, or 14.5 percent, lower than the 234,000 year-earlier average.

FRED puts today's level in striking historical context. Its seasonally adjusted initial-claims series begins in January 1967 and records readings of 197,000 or lower repeatedly in the late 1960s, including 197,000 on Feb. 8 and Nov. 15, 1969, and 182,000 on Sept. 6, 1969. More recently, the Labor Department said a 2022 reading of 187,000 was the lowest since that Sept. 6, 1969 level. The important formulation is therefore that present claims are near historically low nominal levels, not that Sept. 26 set a 57-year record.

The Labor Department's most reliable state-to-state comparison in the Oct. 1 release covers the week ending Sept. 19, because those figures have received the states' second reporting. The largest increases in initial claims were California, up 2,352; Hawaii, up 1,524; New York, up 868; Texas, up 768; and Illinois, up 738. The largest declines were Massachusetts, down 501; Kentucky, down 326; Arkansas, down 265; Puerto Rico, down 124; and Washington, down 120. California and Hawaii supplied no explanatory comments for their increases.

The release also provides advance state numbers for Sept. 26, but the Labor Department explicitly warns that those numbers are not directly comparable with the previous week's state figures. Advance claims are assigned to the state responsible for paying the benefit, while later data identify claimants by state of residence and incorporate work-sharing adjustments. That methodological warning argues against presenting apparent Sept. 26 state changes as clean measures of local layoffs.

Weekly unemployment claims do not provide a national industry breakdown, so sector claims impacts are unspecified in the federal weekly release. Other labor data fill in some context. BLS said August payroll gains occurred in food services and drinking places and local government education, while the information sector lost jobs. Its August JOLTS survey showed about 7.1 million job openings, 5.2 million hires and 1.6 million layoffs and discharges, with layoffs essentially unchanged and little change across industries.

A separate Oct. 1 survey from Challenger, Gray & Christmas showed 43,281 announced U.S. job cuts in September, down 18 percent from August and 20 percent from September 2025. Technology companies announced 10,799 cuts, up 77 percent from August, while food producers announced 7,326. Technology nevertheless coexists with an overall economy showing extremely few actual unemployment-insurance filings; announced cuts and UI claims measure different things and should not be treated as interchangeable.

Taken together, the claims data describe a labor market in which employers remain highly reluctant to dismiss existing workers. They do not, however, show an equally vigorous pace of new hiring. BLS reported that job openings slipped to 7.1 million in August, while payroll employment increased by 162,000 and unemployment remained at 4.1 percent. This "low firing, cautious hiring" combination helps reconcile historically low claims with reports of some job seekers taking longer to find work.

Wages offer a useful check on the overheating argument. Average hourly earnings rose 0.3 percent in August and 3.1 percent over the previous 12 months, according to BLS. That pace indicates continued nominal wage gains, but it does not by itself show the kind of accelerating wage pressure that extremely low claims might otherwise imply.

Federal Reserve officials have increasingly characterized the labor side of their mandate as stable. On Sept. 29, Fed Governor Michael Barr said labor supply and demand appeared to be in "rough balance," with the 4.1 percent unemployment rate close to estimates of maximum sustainable employment. He said risks to the labor market had receded while inflation risks increased. Two weeks earlier, the FOMC unanimously raised its target rate by a quarter percentage point to 3.75 percent to 4 percent, saying job gains had kept pace with workforce growth and unemployment had changed little while inflation remained elevated.

The claims report therefore leans, at the margin, toward the Fed having less reason to ease policy to protect employment. That is an inference rather than an explicit Fed response to the Oct. 1 data. The Fed's September projections put the median end-2026 federal funds rate at 4.1 percent; with the current target range centered at 3.875 percent, that projection is broadly consistent with about one additional quarter-point tightening, but the projections are not a policy commitment.

Private economists made a similar distinction between resilience and overheating. Carl Weinberg of High Frequency Economics told Reuters that high energy and materials costs could eventually pressure employers to shed workers but said the claims data showed no such effect yet. Stephen Stanley of Santander said sustained historically low layoffs could create an overheating risk that the FOMC must watch, though he did not think the economy had reached that point. Samuel Tombs of Pantheon Macroeconomics, meanwhile, expects layoffs to rise next year as tighter financial conditions and companies' adoption of artificial intelligence take hold.

The strongest caveat is revision risk. Initial weekly figures are advances, and states report the data again the following week. In the Oct. 1 release, the Sept. 19 initial-claims estimate moved from 197,000 to 198,000, while Sept. 12 continuing claims were revised from 1.719 million to 1.712 million. The Labor Department says these revisions reflect subsequent state reports and, for the broader historical series, updated unadjusted data and seasonal factors.

Seasonal adjustment also matters unusually much for weekly claims. Holidays, school calendars, weather and other recurring events can produce sharp week-to-week movements. BLS supplies annual seasonal factors to the Labor Department's Employment and Training Administration, and the Labor Department cautions that weekly administrative claims data are difficult to seasonally adjust and therefore subject to volatility. The four-week average of 200,000 is consequently a better short-term trend measure than any single weekly observation.

Finally, claims cover insured unemployment, not every unemployed American. Continuing claims approximate the number of insured unemployed workers claiming benefits, while the BLS unemployment rate comes from a separate household survey and includes people regardless of unemployment-insurance eligibility.