PortfolioAI macro research · September 24, 2026

U.S. Recession Outlook: Growth Cushion Versus Year-End Odds

Employment has not signaled a downturn, but slower output growth leaves the economy more exposed to a demand shock than a single low market quote suggests.

Executive summary

Expansion remains the base case for the next three and six months. August unemployment was 4.1%, and the August Sahm Rule reading was −0.07 percentage points—well below its historical +0.50 warning threshold. Real GDP increased from $24.18 trillion in Q1 to $24.27 trillion in Q2, measured in chained dollars at an annual rate. That gain is positive, but modest: roughly 0.37% quarter on quarter from the displayed levels. The question is whether demand can keep outpacing a weakening margin for error.

Next 3 months

Recession unlikely on the latest labor and output evidence. A sustained jobless-rate rise would challenge this view.

Next 6 months

Expansion favored, with greater sensitivity to weaker hiring, consumer spending and credit availability.

Next 12 months

More uncertain: the 2026 event-market contract does not cover the full period through September 2027.

These horizon assessments are qualitative judgments, not calibrated 3-, 6- or 12-month probabilities. Monthly and quarterly observations below have different publication lags.

Three views of the cycle

The labor trigger has moved away from its threshold

Sahm Rule, percentage points; +0.50 is the historical warning line. Observations through August 2026.

Source: Federal Reserve Bank of St. Louis, SAHMCURRENT. A negative reading does not rule out a future downturn.

The model's last new observation predates September

Recession-probability series, percent; underlying monthly changes through July 2026.

Source: RECPROUSM156N. Daily repeated values reflect carry-forward of the last monthly observation, not new daily estimates.

Real GDP: still rising, more slowly

Quarterly level, trillions of chained dollars at annual rates; Q2 2026 latest.

Source: GDPC1. Changes calculated from rounded displayed levels are not official headline annualized growth rates.

Recession risk scorecard

The observation date matters as much as the reading; a dated weekly indicator cannot establish September conditions.

IndicatorLatest underlying readingWhat it says—and what it does not
Sahm Rule−0.07 pp · Aug 2026Below the +0.50 historical trigger. It detects worsening labor conditions; it is not a forward guarantee.
BCIG8.3 · May 8, 2026Recovered from 4.0 on April 3; too old for a September real-time assessment.
Recession-probability series0.76% · Jul 2026 underlying observationLow model reading, not a September update or comparable to the year-end event contract.
Real GDP$24.27tn · Q2 2026Up from $24.18tn in Q1; continued expansion but only a modest sequential gain.
Unemployment4.1% · Aug 2026Down from 4.4% in December 2025; a persistent reversal matters more than one print.
Money-market fund assets$8.44tn · Q2 2026Up from $8.29tn in Q1; a measure of cash allocation, not a stand-alone recession call.
“Recession” search volume2,202 US searches · Sep 23, 2026Interest in the term, not the number of people losing jobs or a recession forecast.
Polymarket 2026 recession contractAbout 10–11% Yes · Sep 24 search snapshotTradable, changeable year-end contract price; resolution rules, trading liquidity and horizon differ from macro models.

Professional commentary and outlook

The strongest near-term argument against recession is not that traders assign a low price to a year-end event. It is that unemployment has fallen since December while the Sahm Rule has retreated from February's +0.27 to August's −0.07. Recessions can emerge before that lagging trigger trips, however. Its value is in identifying when labor-market damage has become broad enough to validate a deteriorating story.

Output is the more delicate part of the argument. The Q2 real-GDP level rose by $89.2 billion in annual-rate chained dollars from Q1, a smaller step than the $124.7 billion gain in the prior quarter. Two successive negative quarterly changes would be a materially different backdrop, but neither the latest GDP level nor the current labor reading establishes that condition. The July recession-probability model is also reassuring only within its own dated methodology; repeated daily values after July should not be mistaken for fresh September measurements.

Sentiment sits on a different clock. Searches for “recession” totaled 2,202 on September 23, versus 2,078 a day earlier, while the year-end Polymarket contract was around 10–11% on September 24, above the roughly 8.5% quote observed on September 23. Neither move, by itself, overrides the hard data. Search interest measures attention, and a market price is not an official economic forecast. The 2026 contract also expires well before the end of this report's twelve-month horizon.

What would change the assessment: several unemployment readings moving higher together, a Sahm value advancing toward +0.50, and negative sequential real-output growth would strengthen the recession case. Stable employment and another positive GDP quarter would favor continued expansion. For portfolios, stress-test demand-sensitive earnings and refinancing costs rather than treating low year-end odds as permission to ignore downside risk. Retail enthusiasm around AI infrastructure and cloud deals in today's Reddit stock analysis makes the distinction important: a durable contract can still require substantial upfront capital and a healthy customer to deliver shareholder returns.

Data notes: observation periods differ; “latest” refers to each series' own available underlying observation as of September 24, 2026. Event prices are snapshots, not investment advice.