U.S. macroeconomic dashboard · October 1, 2026
Recession Probability Edges Up While Employment Stays Firm
A small rise in the model signal is worth watching, but it has not become a broad contraction in output or jobs.
Executive Summary
The U.S. expansion remains the better-supported near-term case. The Chauvet–Piger smoothed recession-probability series rose from 0.24% in July to 0.62% in August, a meaningful move off a very low base—not a 62% recession call. The latest Sahm Rule reading was −0.07 percentage points in August, well below its 0.50-point historical trigger. Unemployment was 4.1% in August; the latest reported real GDP level, for the second quarter, stood above the first quarter. These readings do not establish a recession, although they cannot rule out a future shock.
Next 3 months
Unlikely, on current evidence. The labor trigger is quiet and output has not rolled over. A sharp employment reversal would change this assessment.
Next 6 months
Not the base case, but more exposed. A persistent rise in unemployment, followed by weaker spending and production, would make the model's uptick more consequential.
Next 12 months
Material uncertainty. No current indicator justifies a precise one-year probability; the year-end event contract is not a 12-month forecast.
Assessment is qualitative, not a statistical forecast. Data have different publication dates; later dates shown by a carried-forward daily series are not new observations.
Three Views of the Cycle
Labor: Sahm Rule versus its 0.50-point threshold
Monthly observations; below the threshold is not a guarantee against recession.
Smoothed recession probability, percent
August's 0.62% remains low even after more than doubling from July's 0.24%.
Real GDP: level, billions of chained dollars
Quarterly observations, not daily GDP. Q2 2026 rose about 0.55% from Q1 (about 2.2% annualized, calculated from displayed levels).
Recession Risk Table
| Indicator | Latest distinct observation | Reading | Interpretation |
|---|---|---|---|
| Sahm Rule (SAHMCURRENT) | Aug. 2026 | −0.07 percentage points | Below the 0.50-point trigger; no labor-based recession signal. |
| Business Cycle Index Growth (BCIG) | May 8, 2026 | 8.3 | Above its zero warning line, but a stale weekly observation; not a current October reading. |
| Chauvet–Piger smoothed recession probability (RECPROUSM156N) | Aug. 2026 | 0.62% | Up from July's 0.24%; still very low. Estimate of recession conditions in August, not next-year odds. |
| Real GDP (GDPC1) | Q2 2026 | $24,408.0 billion, chained dollars | Above Q1's $24,274.4 billion; the level alone says little about coming quarters. |
| Unemployment rate (UNRATE) | Aug. 2026 | 4.1% | Down from 4.2% in June; watch for a sustained upward turn. |
| Money-market fund assets (MMMFFAQ027S) | Q2 2026 | $8.441 trillion | Large liquid balances, but not a direct recession-probability measure or a promise of spending. |
| U.S. searches for “recession” | Sept. 30, 2026 | 1,895 daily searches | Down 13.5% versus roughly seven days earlier; measures attention, not economic output. |
| Polymarket: U.S. recession by end-2026 | Early Oct. 2026 snapshot | About 8–9% “Yes” | Contract-specific year-end odds; not comparable to a smoothed estimate of current conditions or a 12-month forecast. |
Monthly and quarterly series are labeled by the underlying observation period. The smoothed recession series estimates conditions in the observed month; it does not forecast next-year odds. Search attention, traded event prices and realized activity measure different things and should not be averaged together.
Professional Commentary & Outlook
The interesting change is direction, not magnitude. August's smoothed recession estimate more than doubled from July, but at 0.62% it remains a fraction of one percent. Meanwhile the Sahm measure slipped further below its trigger and the unemployment rate held near 4%. If the model's rise reflects genuine weakening, that weakness has yet to appear in the labor readings shown here. The divergence deserves monitoring rather than an all-clear or a recession declaration.
Real GDP supplies a second check: Q2 output was higher than Q1, but quarterly levels are backward-looking and subject to revision. The last BCIG point is from May; treating its 8.3 as a fresh weekly reading would overstate the evidence. Money-market assets topped $8.4 trillion in Q2, which can represent caution, attractive yields or cash management. It is not equivalent to a sudden withdrawal from consumption.
Public anxiety and event prices tell another story. Search interest for “recession” eased over the latest week, while the year-end Polymarket contract put “Yes” near 8–9%. That contract has its own resolution rules and expires in 2026; it is neither a clean 12-month probability nor the same object as the Chauvet–Piger estimate of current recession conditions. More telling than either would be a sustained rise in unemployment accompanied by declines in output and consumer spending.
What changes the call: a Sahm reading approaching 0.50, a persistent rise in joblessness, successive falls in real activity, and a renewed rise in forward-looking recession measures together would make a contraction more plausible. Without that convergence, diversified positioning and scenario discipline are more defensible than treating a single low-but-rising model number as a trading signal.
Sources: Federal Reserve Bank of St. Louis (Sahm), Chauvet–Piger smoothed probabilities, BEA real GDP, BLS unemployment, money-market assets, iMarketSignals BCIG, Daily Search Volume, Polymarket. Figures reflect the latest distinct observations available for this October 1 assessment; series may subsequently be revised.