Recession Risk 38/100 — September 22, 2026
US recession risk over the next 90 days is MODERATE, not elevated, because the most reliable near-real-time labor trigger is still clearly off: the Sahm Rule is reported at -0.07 (well below the 0.50 trigger) and initial claims just printed 196k (week ended Sep 12, reported Sep 17, 2026), the lowest since mid-July. However, the Fed turned incrementally more restrictive on Sep 16, 2026, hiking 25 bps to a 3.75%–4.00% target range and signaling inflation concerns may require further tightening. Growth does not look recessionary in real time—Atlanta Fed GDPNow still points to solid Q3 momentum (4.4% SAAR as of Sep 10, 2026)—but a cluster of leading/cyclical signals (temporary help contraction, weak freight, soft housing, low savings) argues for caution and a higher probability of a downside surprise than markets are pricing.
Recession Risk Score: 38/100 — MODERATE (+4 vs 30 days ago)
Today’s Recession Risk Score is 38/100 (MODERATE), up +4 points versus 30 days ago (34 → 38). The increase is not being driven by a classic “imminent recession” labor breakdown—initial claims are still very low and the Sahm Rule remains well below trigger—but by a late-cycle mix of tightening policy, softening cyclical leaders (housing, freight, temp help), and frothy market pricing. In short: the economy is not rolling over in real time, yet downside asymmetry is rising because leading and financial-vulnerability indicators are clustering on the wrong side.
Score Trend — Last 30 Days
The last 30 days show a stepwise grind higher rather than a straight-line deterioration. The score started at 34 on 2026-08-23 and ends at 38 on 2026-09-22 (+4). The range was tight—min 34, max 38, with an average of 36—which matters: we’re not seeing a panic spike, we’re seeing persistent pressure.
The shape is best described as plateau-with-jumps. Over the last 10 readings, the index repeatedly toggled between 34 and 38 (e.g., 9/14, 9/18, 9/22 at 38), suggesting the model is responding to a small set of “on/off” drivers (policy impulses, rate-curve shifts, market/lending conditions) rather than a broad collapse in coincident data. That pattern is consistent with a late-cycle environment where recession risk is conditional—it rises quickly if labor turns, but it can also mean-revert if housing stabilizes and financial conditions remain loose.
Key Drivers
-
Labor trigger remains clearly “off” (big offset to recession risk)
- Initial claims: 196k (week ended Sep 12, reported Sep 17, 2026)—the lowest since mid-July, signaling layoffs are still rare. (dol.gov)
- Sahm Rule: -0.07 (safe) — far from the 0.50 recession trigger.
- Net: This is the single strongest argument for MODERATE (not HIGH) risk.
-
Fed policy just tightened—incremental headwind to growth
- The FOMC raised the target range by 25 bps to 3.75%–4.00% on Sep 16, 2026, explicitly emphasizing inflation is still elevated. (federalreserve.gov)
- A hike this late in the cycle typically works through housing, credit availability, and business capex with a lag—exactly where our leading indicators are already soft.
-
Growth tracking is solid in “nowcast” terms, but composition is mixed
- Atlanta Fed GDPNow showed 4.4% SAAR for Q3 (as of Sep 10), then updated to ~5.1% SAAR (as of Sep 16–17)—hardly recessionary. (atlantafed.org)
- However, GDPNow commentary also noted residential investment nowcasts moving more negative after housing starts—consistent with our housing-warning cluster. (atlantafed.org)
-
Cyclical “canaries” are weakening despite strong headline PMI
- Temporary Help Services: 2520K (DANGER) — temp help is a classic early labor-cooling leader.
- Freight Transportation Index: -0.3 (DANGER) — goods-side softness persists, consistent with a growth downshift.
-
Financial conditions and credit spreads remain supportive (offset)
- Chicago Fed NFCI: -0.56 (SAFE) — loose conditions.
- HY OAS: ~270 bps (SAFE) — tight spreads imply investors are not pricing broad stress.
- SLOOS: easing in your tracker — suggests credit transmission is not (yet) biting.
-
Household buffers look thin (amplifier risk)
- Personal savings rate: 3.0% (WARNING) — very low cushion if employment weakens.
- Credit card delinquency: 2.9% (WATCH) and debt service: 11.2% (WATCH) — not a crisis, but directionally unfavorable.
Category Breakdown
Using today’s CATEGORY BREAKDOWN counts:
-
Primary Indicators: 3 safe / 4 watch / 2 danger
Mixed: labor triggers are still safe, but the presence of “danger” in primary/leading labor (temp help) keeps the base-case vulnerable to a fast turn. -
Secondary Indicators: 2 safe / 0 watch / 1 danger
Secondary signals are not confirming recession broadly, but the “danger” reading is enough to prevent complacency. -
Housing & Construction: 0 safe / 0 watch / 2 danger
Housing is the cleanest weak pocket in the dashboard; without stabilization here, tighter policy will keep dragging on cyclicals. -
Business Activity: 2 safe / 1 watch / 0 danger
Business activity isn’t recessionary in aggregate—this is the “why not HIGH risk” category. -
Consumer Credit Stress: 1 safe / 2 watch / 1 danger
Stress is creeping, not breaking—watch for spillover from low savings into delinquencies. -
Market Signals: 6 safe / 3 watch / 5 danger
Markets are simultaneously calm (VIX, spreads) and stretched (valuation/fiscal/ratio signals), which often precedes volatility spikes rather than predicting timing. -
Liquidity: 0 safe / 1 watch / 2 danger
Liquidity signals are increasingly thin; the system can function fine—until it doesn’t—so this category is a tail-risk amplifier. -
Real-Time / High-Frequency: 0 safe / 1 watch / 1 danger
High-frequency data is not screaming recession, but it’s not giving an “all clear” either.
Biggest Movers
Top 5 by |7-day % change| (and what they imply):
-
ON RRP Facility ($5B): -74.9% (7D)
Liquidity normalization / depletion is risk-amplifying (confirmatory). It doesn’t cause recession alone, but it can magnify funding shocks. -
Yield Curve (2s30s) (0.81): +11.2% (7D)
Steepening is often interpreted as positive, but post-inversion steepening can be late-cycle. Confirmatory for “late cycle,” not necessarily immediate recession. -
NY Fed Recession Probability (0.9%): -10.6% (7D)
This move is contradictory (improving)—probability down suggests the curve-based model is seeing less recession risk. -
VIX (15.4): -9.4% (7D)
Contradictory (improving)—market volatility is falling into complacency, which tends to reduce near-term stress while increasing fragility. -
Yield Curve (2s10s) (0.25): +2.9% (7D)
Mild steepening is confirmatory of the transition from inversion, but again: historically, the “danger zone” can be the period after inversion ends.
90-Day Indicator Trends
Your 90-day histories show a critical theme: coincident-ish activity is stable to improving, but cyclical leaders and buffers are weak.
Labor / labor-leading
- Initial claims improved from 226k (Jun 24) → 215k (late Jun/early Jul) → 196k today (reported Sep 17 per today’s snapshot). That’s a clear downward trend—not what you see heading into imminent recession. (dol.gov)
- Sahm Rule fell from 0.10 (Jun 24) → 0.07 (early Jul) → -0.07 now (today’s reading). Directionally, this is disinflationary/labor-stable, not recession-triggering.
Output / income
- Industrial Production rose from 102.6 (Jun 24) → 103.1 now (SAFE). That’s modest, but it is not contraction.
- Real personal income ex transfers rose from $16.5T (Jun 24) → $16.6T (late Jun onward), holding steady into July (WATCH). Stable income helps explain why consumer spending has held up.
Housing (weak)
- Housing starts improved from 1177K (Jun 24) to 1275K now, but remains flagged as below trend and weak.
- Building permits slipped from 1413K (Jun 24) → 1394K now, reinforcing forward softness (permits lead starts).
Financial conditions / credit
- NFCI remained loose (~ -0.51 to -0.56) across the history window—supportive risk backdrop.
- High yield spreads stayed tight: 265 bps (Jun 24) → ~270 bps recently. This is not a market pricing recession.
- Yield curve 2s10s stayed positive in the history shown (~0.34 → ~0.36), consistent with a post-inversion normalization regime, but not an “easy money” signal.
Household vulnerability (thin buffers)
- Personal savings rate rose from 2.6% (danger) → 3.0% (warning), but that is still extremely low—meaning households have less cushion if unemployment rises.
- Credit card delinquencies held around 2.9%—not accelerating in the 90-day window shown, but elevated enough to matter if rates stay high.
Bottom line from the 90-day lens: the economy is not sliding into recession mechanically (labor/output aren’t confirming), but the margin of safety is shrinking (housing + buffers + late-cycle policy).
Stock Screener Signals
Today’s screener is dominated by “value dividend” flags—ARCC, AIG, BBY, FNF, HMC, T, LTM, BCE—plus two “oversold growth” names (CHTR, TLK) with low RSI (28–30). That combination typically appears when markets are doing two things at once: (1) seeking cash-flow durability, and (2) selectively bottom-fishing in beaten-up growth rather than buying broad cyclicality.
Interpreting the macro signal: this is consistent with a late-cycle “barbell” posture—investors want dividends and balance-sheet resilience, but they are not abandoning risk entirely (NASDAQ near highs in your dashboard). In recessionary setups, you usually see a more uniform shift toward defensives and higher volatility; here, the VIX is low and spreads are tight, which supports the view that this is not an imminent downturn—but also raises the odds that any disappointment (jobs, inflation, or credit event) is underpriced.
One note: several displayed dividend yields look mechanically abnormal (triple-digit yields), which often indicates data/vendor quirks (special dividends, annualization artifacts, or stale price/dividend fields). Treat the classification (value/dividend vs oversold growth) as more informative than the raw yield numbers.
Latest Economic Developments
-
Fed tightened on Sep 16, 2026: the FOMC raised rates to 3.75%–4.00%, with the statement emphasizing inflation remains elevated and policy is aimed at returning inflation to 2%. (federalreserve.gov)
This matters for recession risk because it raises the probability that weakness in housing and interest-sensitive credit becomes more persistent into Q4. -
Jobless claims fell to 196,000 (Sep 17 release), reinforcing that layoffs remain limited and the labor market is still absorbing workers. (dol.gov)
For recession forecasting, claims are among the best near-real-time triggers—this print argues against a near-term recession call. -
GDPNow has been strong in September: Q3 tracking moved from 4.4% (Sep 10) to ~5.1% (Sep 16–17), signaling real-time growth momentum remains solid even as housing remains a drag in the model’s subcomponents. (atlantafed.org)
-
Conference Board LEI edged down: the LEI fell -0.1% in Aug 2026 to 99.5, after +0.2% in Jul, marking a mild softening but not a deep, persistent deterioration. (conference-board.org)
Near-Term Outlook (Next 30 Days)
Base case for the next month: growth holds up, recession risk stays MODERATE, but the distribution is fat-tailed because policy is restrictive and leading cyclicals are weak.
Catalysts that could push the score higher (worse):
- Claims trend reversal: a sustained move back above roughly the low-200k range for several weeks would be the first clean “turn.”
- Unemployment drift + Sahm acceleration: if unemployment rises enough to push the Sahm Rule meaningfully upward toward 0.50, the score would likely re-rate quickly.
- Credit transmission: watch HY spreads—your own thresholds (≈ 350–400 bps) are reasonable markers for when “tight spreads” stop being a stabilizer.
Catalysts that could push the score lower (better):
- Housing stabilization: permits/starts stop deteriorating and mortgage-sensitive demand steadies.
- A calmer inflation narrative that lets the Fed pause without jawboning further tightening.
On the calendar, markets will focus on the next weekly claims prints, the next GDPNow updates, and the usual month-end macro cluster (income/outlays, inflation measures, and business surveys).
Long-Term Outlook (3-6 Months)
Over a 3–6 month horizon, the macro setup looks like late-cycle resilience with rising policy and fragility risk. The positive side of the ledger is clear: claims are low, GDP nowcasts are strong, financial conditions are loose, and credit spreads are tight. Those are not recession conditions.
But the negative side is increasingly structural:
- Housing and interest sensitivity remain weak and are most exposed to further tightening.
- Household buffers (savings) are thin, increasing the chance that a labor wobble translates into spending pullback faster than in prior cycles.
- Market valuation / liquidity signals in your dashboard are stretched; these don’t time recessions, but they do raise the probability of a destabilizing drawdown that tightens financial conditions rapidly.
Historically, the danger in a post-inversion environment is not that “the curve is positive so we’re safe,” but that the economy can remain fine until labor turns—then the feedback loop (income → spending → employment) can tighten quickly. For now, the best read remains: recession risk is conditional rather than baseline.
What to Watch
Hard thresholds / triggers
- Sahm Rule: watch for moves toward 0.50 (recession trigger).
- Initial claims: watch for a persistent reversal from sub-200k/low-200k into a sustained uptrend.
- HY OAS: watch >350–400 bps as a regime change from “benign” to “credit caution.”
- Housing: permits and starts—continued deterioration into Q4 would keep the leading edge weak.
Event-driven
- Next GDPNow update cadence (Atlanta Fed’s schedule shows the next update on Sep 25, 2026). (atlantafed.org)
- Fed communication after the Sep 16 hike—whether officials reinforce “more tightening” versus “data dependent.” (federalreserve.gov)