US unemployment exceeds 7%

Where things stand

• Latest value: September 2026 U-3 is 4.2%, published October 2, 2026. The 2026 path so far is 4.3, 4.4, 4.3, 4.3, 4.3, 4.2, 4.1, 4.1, 4.2. Source: BLS series LNS14000000, pulled from the BLS flat file https://download.bls.gov/pub/time.series/ln/ln.data.1.AllData.
• What YES needs: a monthly reading of 7.1% or more, about +2.9 points. Each quarter-end can only include data through the second month of that quarter (for example, November data for December 31). The window is therefore October 2026 to November 2031, 62 unknown months.
• Labor market: low hiring but also low layoffs. September added 29k jobs and July and August were revised down. Initial jobless claims are about 199k, near a 57-year low (Reuters, October 15).
• Macro risks are elevated:
  • Brent oil touched about $110 on October 16 as Iran stepped up attacks around the Strait of Hormuz (CNBC).
  • The Fed under Warsh raised rates to 3.75–4% on September 16 and signalled more hikes.
  • Mortgage rates are above 7%.
  • Deloitte notes the personal saving rate has been at or below 3% since April 2026.
  • The IMF and others warn the AI investment boom could end.
• Official forecasts see no problem: the Fed's September projections have a median of 4.1% for 2026–29, with risks roughly balanced (https://www.federalreserve.gov/monetarypolicy/fomcprojtabl20260916.htm). The Philadelphia Fed's Q3 2026 Survey of Professional Forecasters (prob.xlsx) puts under 0.5% on an annual average of 7.2% or more in any year from 2027 to 2029. These surveys are known to be overconfident in the tails.

Base rates (computed from LNS14000000, 1948–2026)

• Starting from a similar rate: among months with unemployment between 3.9% and 4.6%, the share where some month above 7.0% followed within 62 months was:
  • 54% using data from 1948 on;
  • 40% from 1960 on;
  • 43% from 1984 on. These months group into only about 10 independent episodes. Since 1993, about 45% of 62-month windows contained a month above 7%, driven by the 2008 financial crisis and COVID.
• Recession severity: in 7 of the 12 recessions since 1948, unemployment rose 2.9 points or more from the month the recession began; 1960, 1969, 1980, 1990 and 2001 fell short. When the threshold was crossed, it took 2 to 17 months (median about 11).
• Recession frequency: about 0.15 per year since 1945, about 0.10 since 1984, and higher when unemployment is below 5%.

Model

• I simulated recession starts at 20% a year for the next 15 months, then 15% a year. A recession crosses 7% with probability 0.50 for near-term recessions and 0.55 for later ones.
• These are trimmed below history because labor force growth is close to zero with low immigration, so a mild recession would raise the rate less than in the past.
• I used the historical lag distribution and added a small extra chance (rising to about 2% a year) of AI-driven job losses pushing unemployment over 7% outside a normal recession.
• Result: about 32–34% by the end of 2031. I nudged this up to about 37% because the historical base rates (40–55%) and crowd forecasts are higher.
• Near-term values stay tiny: getting above 7% by February 2027 would take a COVID-scale shock.

Market and crowd checks

• Kalshi "Recession in 2027" (two negative GDP quarters): 20%.
• Polymarket "US recession by end of 2027": about 34%, a thin market.
• Blue Chip consensus: 29% chance of recession within 12 months.
• Kalshi's market on how high unemployment will get in 2026 prices above 7% at about 1–3%.
• Deloitte's AI-bust downside scenario has unemployment peaking at 7.2%, right at the threshold.
• Metaculus puts a 10% rate before 2031 at 55%. Kalshi's "before 2030" maximum-unemployment ladder is illiquid and internally inconsistent (it prices above 7% higher than above 6%). I gave both little weight, reading them as tilted toward an AI-disruption view.

Strongest case against my estimate

• Higher: history says a low-unemployment expansion almost always ends in recession within about four years. The oil shock combined with rate hikes, plus possible AI displacement, could make the next downturn severe.
• Lower: professional forecasters and the Fed see the rate flat near 4.1–4.3%. Layoffs are at record lows, and slow labor force growth dampens increases in the rate. Expansions since 1984 have been long.

On balance I put about 37% on a reading above 7% by the end of 2031, with the hazard per quarter rising through 2028 and then roughly flat.

Question. By the last day of each quarter, has BLS reported a US unemployment rate exceeding 7% for any month?

Definitions.

• US unemployment rate: the monthly, seasonally adjusted official U-3 unemployment rate from the Current Population Survey, BLS series LNS14000000: unemployed people as a percentage of the civilian labor force aged 16 and older in the 50 states and District of Columbia.
• Exceeding 7%: a published rate strictly greater than 7.0%, meaning at least 7.1% at BLS's standard one-decimal precision. A rate of exactly 7.0% does not count; do not infer greater precision from underlying counts.
• Eligible month: September 2026 or later, under the Observation period convention.

Resolves YES if BLS publishes a qualifying rate for an eligible month on or before the quarter-end. Either an initial release or an officially revised value counts. Apply the Resolution window convention: a subsequent downward revision does not undo YES.

Resolves NO if no qualifying publication exists by the quarter-end. Unadjusted rates, U-1, U-2, U-4 through U-6, state or subgroup rates, forecasts, and revisions solely to months before September 2026 do not count. No attribution to AI is required.

Resolving source. BLS's monthly The Employment Situation, Table A-1 (total unemployment rate) or Table A-15 (seasonally adjusted U-3), and the official LNS14000000 series, including dated BLS revisions. BLS publications only, a narrower source requirement than Public confirmation (A2). Apply Resolution date (A3), using the publication date of the first qualifying value, not the month measured. Apply the Resolution window and Observation period conventions.