Recession Risk 34/100 — September 25, 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 matters: risk has fallen by 4 points versus 30 days ago (38 → 34). The “contained risk” verdict is still anchored by the non‑triggered Sahm Rule (-0.07) and very low initial claims (~196k)—two of the cleanest near-term recession tripwires. At the same time, the score is not low because several leading-cycle components—temporary help, housing, consumer confidence, and goods/transport proxies—continue to flash caution.

Score Trend — Last 30 Days

The 30-day window (2026-08-26 → 2026-09-25) shows a gentle downshift in risk: Start 38, End 34 (Δ -4), with a Min 34 / Max 39 / Avg 36. That’s not a collapse in risk—more like a cooling from “edge-of-elevated” into “mid‑moderate.”

The shape is best described as mean-reverting with recurring spikes. In the last 10 readings, the score repeatedly jumped to 38–39 (Sep 18, Sep 22–24) before snapping back to 34. That pattern usually means the macro backdrop is stable enough to prevent a sustained deterioration, but fragile enough that one or two data points (rates, oil, a hot activity print, or risk-off positioning) can temporarily push “recession chatter” higher without confirming follow-through in labor and credit.

Key Drivers

1. Labor-market “hard stop” indicators remain calm (contain risk)

   • Initial jobless claims: ~196k (recent weekly print), consistent with limited layoff pressure and continued expansion. (apnews.com)
   • Sahm Rule: -0.07 (SAFE) — nowhere near the +0.50pp trigger, keeping “imminent recession” probability capped.

2. The yield curve has re-steepened (reduces classic inversion signal)

   • 2Y ~4.87% vs 10Y ~5.18% on Sep 24, 2026 → 2s10s = +0.31pp. (treasuryratewatch.com)
   • In practice: steepening often reflects less immediate policy-tightening stress (or rising term premium), even if front-end yields remain restrictive in level terms.

3. Leading indicators are soft—but not collapsing

   • The Conference Board LEI fell -0.1% in Aug 2026 to 99.5, following +0.2% in July—mild drag, not a broad plunge. (conference-board.org)
   • This aligns with “sub‑trend growth” rather than a fast break into contraction.

4. The Fed just tightened again (policy error risk rises into year-end)

   • Multiple credible summaries indicate the Fed raised rates 25 bps on Sep 16, 2026 to 3.75%–4.00%, reinforcing inflation persistence concerns and raising late-cycle accident risk. (fedpolicy.org)

5. Late-cycle fragilities keep tail risk alive

   • Temporary help services: DANGER (a classic early warning that firms are reducing flexible labor before cutting core headcount).
   • Housing: permits/starts are weak on your dashboard; housing tends to lead turns.
   • Consumer sentiment: your reading is 55.2 (WARNING)—still subdued versus long-run norms, implying fragile discretionary demand (even if month-to-month prints can bounce). (data.sca.isr.umich.edu)

Category Breakdown

• Primary Indicators: 3 safe / 4 watch / 2 danger
  Mixed: the “big recession triggers” aren’t firing (claims/Sahm), but the watch/danger share says the expansion is aging and susceptible to shocks.

• Secondary Indicators: 2 safe / 0 watch / 1 danger
  Thin coverage here, but the split argues for contained near-term risk with one meaningful weak point.

• Housing & Construction: 0 safe / 0 watch / 2 danger
  Housing is the clearest cyclical stress pocket; prolonged weakness here can leak into employment, durables demand, and regional credit.

• Business Activity: 2 safe / 1 watch / 0 danger
  Still more “slow growth” than “shutdown,” consistent with a soft-landing posture.

• Consumer Credit Stress: 1 safe / 2 watch / 1 danger
  Credit is not breaking, but it’s no longer pristine; watch-list items (delinquencies, debt service) suggest household buffers are thinner.

• Market Signals: 6 safe / 3 watch / 5 danger
  Markets are sending a split message: risk appetite remains high (tight spreads, low vol), but valuation/fiscal/ratio-style signals keep flashing late-cycle excess.

• Liquidity: 0 safe / 1 watch / 2 danger
  Liquidity is a growing “silent driver” risk—especially with cash-facility dynamics shifting quickly.

• Real-Time / High-Frequency: 0 safe / 1 watch / 1 danger
  High-frequency data are not screaming recession, but they’re no longer cleanly expansionary.

Biggest Movers

1. ON RRP Facility ($5B): -94.4% (7D)
   Contradictory/ambiguous: falling RRP can signal liquidity being absorbed elsewhere (or shifting cash-management preferences). In late cycle, rapid facility depletion can reduce “shock absorbers,” nudging risk up, not down.

2. NY Fed Recession Probability (0.9%): -88.4% (7D)
   Contradictory (improving): a sharp decline would normally imply lower modeled recession odds—but given the magnitude, treat as “model/inputs moving” rather than a fundamental all-clear.

3. Housing Starts (1275K): +21.2% (7D)
   Contradictory (improving): a bounce helps the growth outlook at the margin, but the category still screens as danger overall—one pop does not fix a trend.

4. Yield Curve (2s10s) (0.31): +13.9% (7D)
   Confirmatory (improving): further steepening reduces the inversion-era warning signal and typically aligns with lower near-term recession risk.

5. Yield Curve (2s30s) (0.81): +10.5% (7D)
   Confirmatory (improving): a more normal curve shape supports the “moderate, not elevated” near-term call.

90-Day Indicator Trends

Using your 90-day history window (late June → mid/late July snapshots provided) plus today’s readings, the dominant story is stable coincident conditions with worsening leading-cycle texture.

• Industrial production (SAFE): ~102.6 → 103.1 (today). That’s a modest improvement and supports the “no imminent recession” base case.

  • 90 days ago (late June): ~102.6
  • 60 days ago: ~102.6 (flat cluster)
  • 30 days ago: not explicitly listed in the history block, but today’s level is clearly above the June/early‑July plateau.

• Initial claims (SAFE): from ~215k (late June) to ~196k (today)—a meaningful downshift, inconsistent with a recession start.

  • Late June: ~215k
  • Mid-July: ~208k
  • Today: ~196k

• Yield curve (2s10s) (WATCH): consistently positive in your historical block (~0.31 → ~0.42 by mid‑July) and ~0.31 today. Net: not inverted in this window; curve risk is more about level (tight policy) than shape.

• Financial conditions (NFCI) (SAFE): hovering around -0.52 to -0.56 across the window—loose conditions, consistent with resilient risk appetite and supportive for growth (until it isn’t). (ycharts.com)

• Credit spreads (HY OAS) (SAFE): ~278 bps late June → ~270 bps mid‑July → ~270 bps now (tight). Tight spreads are a major reason the model won’t price high recession odds today.

• Household fragility indicators (WATCH/WARNING):

  • Personal savings rate ~3.0% (WARNING) is persistently low in your window—this is less about timing a recession and more about asymmetry: shocks propagate faster when buffers are thin.

• Temp help (DANGER): stuck in danger in the historical block and still danger today—this is the cleanest “yellow/red” lead signal in your dashboard. When temp help declines persistently, it often precedes broader labor softening by months.

Netting it out: over ~90 days, the data mix looks like slow-but-still-expanding coincident activity, with rising vulnerability if labor finally turns (claims trend up, unemployment rises, Sahm accelerates).

Stock Screener Signals

Today’s screener is dominated by value/dividend flags: ARCC, AIG, BBY, FNF, HMC, T, BCE plus a couple oversold growth names (CHTR, TLK) with low RSIs (notably CHTR RSI ~28). In macro terms, that blend usually points to a market trying to do two things at once:

1. Clip carry and hide in cash-flow
   Value/dividend selection tends to rise when investors believe growth will decelerate but not necessarily collapse—consistent with today’s 34/100 moderate reading and the “late-cycle but not recession” posture.

2. Selective mean reversion in beaten growth
   Oversold growth flags suggest pockets of the market are already pricing tighter financial conditions and slower demand. If the macro stays stable (claims and spreads remain benign), those oversold names often lead short, sharp rallies. If credit/labor cracks, they become “value traps.”

One important caveat: the yields shown (e.g., ARCC 1002%, BBY 654%) are not economically plausible as actual dividend yields; treat them as data artifacts rather than investable signals. The style factor message—defensive/value + selective oversold growth—is still useful.

Latest Economic Developments

• Claims remain very low: The latest weekly initial claims print around 196,000, the lowest since mid‑July per the reporting, reinforcing that layoffs remain rare. (apnews.com)
• The Fed tightened on Sep 16, 2026: available summaries indicate a 25 bp hike to 3.75%–4.00%, with messaging that inflation persistence is still a concern—this is the main “macro risk injector” for Q4. (fedpolicy.org)
• LEI softened modestly: The Conference Board reported the LEI down -0.1% in August 2026 to 99.5, consistent with slowing momentum rather than a sharp downturn. (conference-board.org)
• The curve is upward sloping again: On Sep 24, 2026, the 10Y (5.18%) exceeded the 2Y (4.87%) by +0.31pp, a meaningful change versus inversion regimes that historically signaled recession risk. (treasuryratewatch.com)

Taken together: the “now” economy still looks okay (claims, spreads, production), but the “next” economy depends heavily on whether the post‑hike period triggers a delayed break in labor, housing, and confidence.

Near-Term Outlook (Next 30 Days)

Base case for the next month is sub-trend growth with stable labor, keeping the score anchored in the low-to-mid 30s unless we see a coordinated deterioration in claims + unemployment + spreads.

Catalysts most likely to move the score quickly:

• Labor: any sustained move in initial claims (e.g., a multi-week climb toward the mid‑200k+ range) or a meaningful unemployment uptick would pull the Sahm Rule closer to the trigger and could push the score into the 40s rapidly.
• Credit: HY OAS is tight now; a sudden widening would be a high-signal warning that markets are pricing default/recession risk.
• ISM/PMI + housing: if weak housing readings persist and manufacturing employment rolls over, the “yellow lights” could become a more coherent slowdown narrative.

On the calendar, the October 22, 2026 LEI release is a key checkpoint for whether leading indicators are merely drifting or beginning to deteriorate more diffusely.

Long-Term Outlook (3-6 Months)

The 3–6 month outlook is best described as late-cycle slow growth with fattening downside tails.

• The “soft landing” case remains credible because:
  claims are low, financial conditions are loose, and spreads are tight—the usual recession accelerants are not present.
• The “accident” case remains plausible because:
  policy tightened again on Sep 16, housing is weak, temp help is contracting, and consumer buffers (savings) are thin—conditions that can flip quickly if the labor market turns from stable to nonlinear deterioration.

Historically, expansions don’t die from “bad sentiment” alone—they break when employment and credit join the weakness. Right now, the model’s message is: watch the handoff from leading indicators (already soft) into coincident indicators (still okay). If that handoff occurs, the score won’t drift—it will jump.

What to Watch

Hard thresholds / tripwires

• Sahm Rule: watch for movement toward +0.50pp (today: -0.07). A rapid rise is a regime shift.
• Initial claims: trend matters more than one print—watch for a sustained uptrend over 4–6 weeks.
• HY credit spreads (OAS): a move from ~270 bps to >350–400 bps would be a classic “risk-off + growth scare” confirmation.
• Housing: permits/starts need stabilization—not just a one-week bounce—to reduce recession tail risk.

Events

• Next major sentiment and activity reads (ISM prints, confidence surveys) for confirmation of whether weakness is contained to housing/goods or spreading.
• Fed communication: whether officials signal additional hikes or stress “higher for longer,” which would raise policy-error probability.

Sources

• apnews.com
• apnews.com
• conference-board.org
• bls.gov
• dol.gov
• fedpolicy.org
• ey.com
• treasuryratewatch.com
• sca.isr.umich.edu