PortfolioAI macro research · September 30, 2026
U.S. Recession Watch: Attention Climbs, Labor Holds
Search interest is rising, but the latest employment and output readings still point to expansion rather than an imminent downturn.
Executive Summary
The central distinction is between fear of a recession and evidence of one. The August unemployment rate was 4.1%, unchanged from July, while the August Sahm Rule indicator was −0.07 percentage points, well below its 0.50-point historical trigger. The latest available real GDP reading, for the second quarter, was $24.408 trillion at an annual rate in chained 2017 dollars, up from $24.274 trillion in the first quarter (roughly 2.2% annualized quarter-on-quarter growth). Neither series describes a broad contraction today.
Three months: recession is not the base case on the available hard data. Six months: risk is more open-ended because employment can turn after output and consumer demand soften. Twelve months: the present readings are not a reliable all-clear; watch whether a sustained rise in unemployment accompanies weaker real activity. These are qualitative assessments, not calibrated probability forecasts. A Polymarket contract pricing a U.S. recession by the end of 2026 at about 9% covers a different horizon and uses contract-specific resolution rules; it should not be treated as the odds for the next 12 months.
Labor-market buffer
Monthly observations through August 2026. Sahm Rule trigger: 0.50 percentage points; different units use separate axes.
Real output still expanding
Quarterly real GDP levels, billions of chained 2017 dollars, through Q2 2026. Repeated daily values in an interpolated series are not new GDP releases.
Recession Risk Dashboard
| Indicator | Latest available reading | Interpretation |
|---|---|---|
| Sahm Rule (SAHMCURRENT) | −0.07 pp · Aug 2026 | Below the 0.50 pp recession-onset threshold; a negative reading means the current three-month unemployment average is below its prior-year low. |
| Unemployment (UNRATE) | 4.1% · Aug 2026 | Unchanged from July and below December 2025's 4.4%; a persistent reversal would matter more than one monthly print. |
| Smoothed recession probability (RECPROUSM156N) | 0.76% · latest carried reading on Sep 30, underlying update Jul 2026 | Low retrospective model signal, not a forward 12-month recession probability. |
| Real GDP (GDPC1) | $24.408 trillion · Q2 2026 | About 2.2% annualized growth from Q1 on the displayed chained-dollar levels; quarterly and lagged, not a September activity estimate. |
| Business Conditions Index (BCIG) | 8.3 · May 8, 2026 | Rose from 4.0 on April 3 in the available weekly series; too old to treat as a September nowcast. |
| Money market fund assets (MMMFFAQ027S) | $8.44 trillion · Q2 2026 observation | Up from $8.29 trillion at the prior quarterly observation. Cash balances are a positioning/liquidity measure, not a recession trigger. |
| “Recession” searches (U.S.) | 2,162 daily searches · Sep 29, 2026 | About 64.9% above the reading roughly 30 days earlier, yet 61.6% below a year earlier. Attention is a sentiment gauge, not output data. |
| Polymarket year-end contract | About 9% “Yes” · Sep 30 snapshot | Tradable, changing contract price subject to its own resolution criteria; not comparable to the smoothed historical probability above. |
Professional Commentary & Outlook
The near-term recession thesis needs a bridge from anxiety to measurable deterioration. A jump in recession searches can signal heightened concern, but the labor data have moved in the opposite direction: unemployment eased from 4.4% in December to 4.1% in August, while the Sahm measure fell from 0.35 to −0.07. Real output also rose in the latest available quarter. That combination argues against calling a recession based on sentiment or event-contract pricing alone.
The longer-horizon vulnerability is timing. GDP is reported with a lag, the smoothed probability series carries a July underlying reading, and the available business-conditions and money-fund observations are older still. A change in hiring and layoffs could therefore alter the outlook before these slower series fully reflect it. The decisive test is whether the unemployment three-month average begins to climb toward the Sahm threshold and real activity loses momentum—not whether a single headline generates more searches.
For portfolio construction, keep defensive liquidity and position sizing tied to portfolio needs rather than treating a 9% year-end contract as a trading signal. High-quality bonds or cash can buffer a growth shock, but their duration and reinvestment risks differ. Equities with dependable cash flows may be more resilient than highly cyclical balance sheets if labor weakens; if hiring steadies and real output expands, an indiscriminate defensive rotation risks missing the expansion. Reassess after each new employment and GDP release instead of extrapolating stale weekly or quarterly figures into September.