Recession Risk 34/100 — September 3, 2026

Recession Risk Score: 34/100 — MODERATE (-4 vs 30 days ago)

Today’s Recession Risk Score is 34/100 (MODERATE), and the direction of travel has been down over the last month (-4 points vs 30 days ago). The core reason the score is not higher is that the highest-signal, high-frequency labor triggers remain unconfirmed: the Sahm Rule is still untriggered and initial jobless claims are still historically low. At the same time, several late-cycle fault lines remain visible—especially temporary help employment deterioration, weak household confidence, and system fragility (bank unrealized losses + fiscal/interest-expense constraints) that can amplify a slowdown into a sharper event if liquidity tightens.

---

Score Trend — Last 30 Days

Over the last 30 days (window 2026-08-04 → 2026-09-03), the score moved from 38 to 34 (Δ: -4). The range has been contained but meaningful: min 34 / max 42, with an average of 36 across 31 samples—a classic “moderate-risk corridor,” not a panic regime.

The shape of the last 10 readings matters: we’ve seen repeated dips to 34 (Aug 29, Aug 31, Sep 1, Sep 2, Sep 3) interrupted by brief reversion to 38 (Aug 30). That pattern reads as mean reversion lower—risk is stabilizing at a reduced level rather than accelerating higher. In plain English: the economy is slowing at the margins, but it isn’t cascading in the data that typically precedes near-term recession calls (claims spike, credit spreads blow out, broad hiring contraction).

---

Key Drivers

1. Labor market “hard trigger” remains unconfirmed (Sahm Rule safe)

   • Sahm Rule: -0.03 (SAFE) vs 0.50 trigger.
   • This is the biggest single reason the model won’t print “high risk” today. In most modern cycles, a true 90-day recession-risk surge requires either a Sahm-style unemployment acceleration or a claims breakout.

2. Layoffs are still low; claims are consistent with a non-recessionary labor market

   • Initial jobless claims: ~203K (SAFE)—still in the “tight labor” zone.
   • Recent reporting confirms claims around 203,000, reinforcing that layoffs remain limited. (apnews.com)

3. Yield curve normalization is a near-term recession-risk reducer

   • 2s10s: +0.40 (WATCH) and 2s30s: +0.88 (SAFE)—both positively sloped.
   • A re-steepened curve tends to reduce near-term recession odds, even though the memory of prior inversion keeps medium-term risk non-zero.

4. Business cycle leading composites aren’t flashing a classic contraction

   • Your dashboard notes Conference Board LEI is not in a persistent contraction profile (and today’s read is SAFE in your system).
   • That aligns with the broader theme: slowing, not collapsing.

5. Manufacturing is expanding, but cooling at the margin (late-cycle deceleration signal)

   • ISM Manufacturing PMI for August 2026 was reported at 54.6, down from 55.6 in July; new orders also cooled (still expansion). (investing.com)
   • This supports “growth continues, but breadth is narrowing.”

6. Asymmetric downside risk: fragility channels remain elevated

   • Bank unrealized losses: $5,155B (WARNING) plus US debt dynamics (Debt/GDP 123% WARNING; National Debt $39.1T DANGER; Interest expense $1,247B WARNING) create a backdrop where a liquidity accident or policy error can turn a mild slowdown into a sharper contraction.

---

Category Breakdown

• Primary Indicators: 3 safe / 4 watch / 2 danger
  Mixed-to-constructive overall. The “watch” bulk (GDP growth, unemployment rate drift, etc.) is consistent with late-cycle cooling, not a near-term recession call.

• Secondary Indicators: 2 safe / 0 watch / 1 danger
  Still mostly stable; the “1 danger” is a reminder that second-order signals can deteriorate quickly if labor rolls.

• Housing & Construction: 0 safe / 1 watch / 1 danger
  Housing remains a key soft spot—consistent with higher-rate lag effects and affordability constraints.

• Business Activity: 2 safe / 1 watch / 0 danger
  This category is still holding up, fitting with the ISM expansion regime even if momentum is easing. (investing.com)

• Consumer Credit Stress: 1 safe / 2 watch / 1 danger
  Not a blow-up, but stress is building (delinquencies, low savings buffer). This is a slow-burn recession transmission channel.

• Market Signals: 7 safe / 2 watch / 5 danger
  Markets are simultaneously “calm” (VIX low, spreads tight) and “priced for perfection” (valuation/GDP extremes, copper/gold stress). This divergence tends to raise crash sensitivity more than it raises baseline recession odds.

• Liquidity: 0 safe / 1 watch / 2 danger
  Liquidity is the sleeper risk. The system can look fine—until it doesn’t—especially with the RRP buffer largely depleted and bank balance-sheet sensitivity high.

• Real-Time / High-Frequency: 0 safe / 1 watch / 1 danger
  Real-time is mixed: layoffs/claims look fine, but freight/temps suggest the goods economy and hiring intent are weaker than headline markets imply.

---

Biggest Movers

1. ON RRP Facility ($525M): +1328.2% (7D) — confirmatory (worsening fragility risk)
   The direction is less important than the message: the cash “shock absorber” is effectively gone, making funding markets more reliant on smooth functioning.

2. Yield Curve (2s30s) (0.88): -21.1% (7D) — contradictory (slightly higher risk vs prior week, but still SAFE)
   It’s still positively sloped, but the weekly compression suggests rate expectations are shifting.

3. Housing Starts (1239K): -19.7% (7D) — confirmatory (worsening growth risk)
   Housing is often an early cyclical pressure point; a sharp weekly decline reinforces the “soft construction” regime.

4. NY Fed Recession Probability (3.6%): +14.8% (7D) — confirmatory (worsening, but level remains low)
   The level is still benign, but the rate of change matters: probabilities rising from a low base can foreshadow broader risk repricing if macro prints weaken.

5. Yield Curve (2s10s) (0.40): -12.8% (7D) (despite +25.9% 1D) — mixed
   The curve is steep/positive, which is supportive, but the weekly move implies investors are actively re-evaluating the Fed path into September.

---

90-Day Indicator Trends

The 90-day history you provided (with daily marks visible from early June) is consistent with a slow-growth, late-cycle profile—no broad collapse, but clear softness in the “early labor” and “real economy goods” signals.

Industrial production: slow improvement (stable-positive)

• Industrial Production moved from 102.5 (Jun 5) to ~102.6 (mid-June) and is 103.0 today (SAFE).
• That’s a small but meaningful upward drift—not recessionary behavior.

Temporary help: structurally weak, still danger

• Temp help services hovered around 2485K–2490K in June (DANGER) and sits at 2505K today (DANGER).
• The level remaining “danger” despite a modest rise indicates the model is treating this as a late-cycle labor leading signal: firms reduce flexible labor first, then slow hiring, then cut core payrolls.

Housing: step-down weakness persists

• Housing starts show a notable deterioration: ~1465K in early June (WATCH) to ~1177K by mid/late June (WARNING), and 1239K today (WARNING).

• That’s a clear regime shift lower; housing is no longer “cooling”—it’s constrained.

• Building permits eased from 1423K (Jun 5) to 1413K by mid-June, and are 1433K today (WATCH)—stabilizing but not re-accelerating.

Household buffer: low savings is a persistent macro vulnerability

• Personal savings rate was 2.6% in June (DANGER) and is 3.0% today (WARNING).
• Improvement is directionally good, but the level remains thin—limited shock absorption if unemployment rises or credit tightens.

Credit & financial conditions: calm on the surface

• High-yield OAS tightened from ~275 bps in early June to around 265 bps (SAFE).
• Chicago Fed NFCI is negative (loose) and remains supportive.
• Translation: the market is not yet pricing recession—so if labor cracks, repricing risk is large.

Valuation/market breadth: “risk-on” persists, raising crash sensitivity

• S&P 500 in your history shows volatility (mid-June dip), while today the level is 7667 (SAFE)—near highs.
• Yet NASDAQ/GDP ratio (DANGER) and Copper/Gold (DANGER) are simultaneously screaming caution. That combination is typical of late-cycle “two economies”: markets priced for productivity/AI capex while goods/industrial signals imply caution.

---

Stock Screener Signals

Today’s screener is dominated by “value dividend” flags (ARCC, AIG, BBY, FNF, HMC, T, LTM, BCE) plus a smaller pocket of “oversold growth” (CHTR, TLK). Two interpretations matter.

First, the clustering in low P/E, high-yield names suggests market participants are increasingly price-sensitive and favoring cash-flow durability over long-duration growth—consistent with a world where the Fed is still debating hikes and real yields are not collapsing. That is not a recession call by itself; it’s a late-cycle positioning tilt.

Second, note that several quoted “yields” in the data look mechanically extreme (triple-digit yields), which in real markets usually means special dividends, data errors, or trailing distributions that won’t repeat. But even with that caveat, the factor signal is still clear: carry and value are being screened as attractive versus expensive momentum, which fits with your model’s “moderate risk, not elevated” macro regime.

---

Latest Economic Developments

Three developments in the past ~48 hours reinforce today’s score as moderate rather than high:

1. Fed Beige Book: modest growth, slight employment gains, moderate price increases

   • The Fed’s Beige Book (Sep 2, 2026) described modest increases in activity, slight employment growth, and moderate price increases, alongside mixed sentiment and uncertainty tied to energy prices and geopolitical conflict. (marketscreener.com)
   • This is the Fed’s “real-time anecdotal” read: it doesn’t scream recession, but it does suggest a cautious, late-cycle environment.

2. NY Fed’s John Williams: higher long yields reflect strength, not inflation panic

   • Williams framed rising long-term yields as more consistent with a solid economy and strong outlook rather than inflation fears, while the market continues to price a meaningful chance of a hike at the Sep 15–16 FOMC. (investing.com)
   • For recession odds, this matters because it keeps the market leaning toward “soft landing / resilience” rather than “hard stop”—even if it raises policy-error risk.

3. Manufacturing expansion continues, but momentum cooled in August

   • ISM Manufacturing PMI 54.6 (Aug) vs 55.6 (Jul), with new orders down—still expansion, but deceleration. (investing.com)
   • This aligns with the model: goods economy is not in freefall, but it’s not accelerating either.

Separately, the labor market narrative heading into Friday is “tepid but not breaking.” ADP signaled softer private hiring ahead of the official payroll report, keeping the market focused on whether labor cooling becomes broad-based. (axios.com)

---

Near-Term Outlook (Next 30 Days)

The next 30 days are likely to be policy-and-labor dominated. The score can move quickly if either (a) labor cracks beyond leading indicators or (b) the Fed signals a renewed hiking bias into visible deceleration.

Key catalysts:

• Sep 4, 2026 — Employment Situation (Aug 2026 data)
  Watch for:

  • Unemployment rate: does it move above your current 4.1% (WATCH) and, more importantly, does the 3-month moving average start rising fast enough to lift the Sahm Rule?
  • Hours worked / payroll breadth: deterioration here often precedes a claims spike.

• Sep 15–16, 2026 — FOMC meeting
  Markets are actively debating hike odds; recent Fed communications have emphasized inflation vigilance. (investing.com)
  If the Fed tightens into slowing real activity, recession odds rise via:

  • corporate hiring freezes,
  • consumer credit tightening,
  • housing re-weakening,
  • refinancing stress.

• High-frequency stress monitors (weekly)

  • Initial claims: a sustained move from ~200K toward the 240K–260K zone would be an early escalation.
  • Credit spreads: HY OAS moving from ~265 bps toward 350–450 bps would signal recession pricing (not currently happening).

---

Long-Term Outlook (3-6 Months)

Over the next 3–6 months, the macro picture is best described as: soft growth with asymmetric downside.

• The “soft growth” part is supported by:

  • positively sloped yield curves (2s10s and 2s30s),
  • contained credit spreads,
  • stable industrial production trend,
  • Beige Book’s “modest growth” characterization. (marketscreener.com)

• The “asymmetric downside” part is supported by:

  • temp help (DANGER) as a labor leading warning,
  • housing weakness (starts in WARNING),
  • thin household buffers (savings rate low),
  • system/fiscal constraints (high debt, high interest expense, large unrealized losses).

Historical parallel framing: in late-cycle slowdowns, recessions often arrive not because baseline growth goes negative immediately, but because a shock + fragility combination forces rapid deleveraging (credit event, funding stress, geopolitical energy spike, or policy mistake). Your liquidity and valuation clusters are exactly the channels that can transmit such a shock faster than traditional coincident indicators would suggest.

---

What to Watch

Hard thresholds (will move the score quickly):

• Sahm Rule: watch for any persistent move toward 0.20–0.30, and especially acceleration toward 0.50.
• Initial claims: a sustained break above 230K followed by 250K+.
• HY OAS: sustained widening above 350 bps would indicate risk repricing.

Key scheduled events (next few weeks):

• Sep 4, 2026 (Fri): BLS Employment Situation (Aug) (dol.gov)
• Sep 15–16, 2026: FOMC decision (rate path / statement tone / press conference)
• Ongoing: watch for follow-on Fed communications that reinforce (or back away from) a hiking bias. (investing.com)

Soft but important confirms:

• Temporary help: further declines would validate a coming payroll slowdown.
• Housing starts/permits: stabilization is needed to prevent housing from becoming a broader employment drag.
• Confidence/sentiment: sustained weakness increases the probability that consumption slows abruptly once labor cools.

---

Sources

• axios.com
• apnews.com
• dol.gov
• economictimes.com
• investing.com
• investing.com
• marketscreener.com
• marketscreener.com
• marketscreener.com
• marketscreener.com
• prnewswire.com
• uol.com.br
• boursorama.com