Recession Risk 38/100 — September 18, 2026

Recession Risk Score: 38/100 — MODERATE (+4 vs 30 days ago)

Today’s Recession Risk Score is 38/100 (MODERATE), up +4 points from 30 days ago (34 on August 19, 2026). The score drifted higher because policy tightened into late-cycle fragility—not because the classic “hard triggers” have flipped. Jobless claims remain exceptionally low (196k for the week ending September 12, 2026) and credit spreads remain tight (HY OAS ~270 bps), which keeps near-term recession odds contained. Net: the economy still looks like slowdown risk rather than imminent contraction, but the balance of risks is moving in the wrong direction.

Score Trend — Last 30 Days

The past month has been a range-bound grind higher, not a straight-line deterioration. The window runs 2026-08-19 → 2026-09-18, with the score moving from 34 → 38 (+4), with a min of 34, max of 38, and average of 36 (31 samples). The pattern reads like a late-cycle regime where “soft data” and select leading indicators worsen, while the labor market and credit remain stubbornly resilient.

The last 10 readings show a two-state dynamic: repeated drops back to 34 followed by quick jumps to 37–38 (notably Sep 10, Sep 14, and Sep 18 at 38). That mean-reverting sawtooth suggests the market/economy is not yet in a self-reinforcing downturn loop; instead, the score spikes are being driven by discrete catalysts—chiefly the Fed’s September 16 hike and persistent weakness in confidence + select leading labor indicators (temp help)—with stabilizers coming from claims, spreads, and still-okay “hard activity” proxies.

Key Drivers

1) Monetary policy shock: Fed tightened on September 16 (restrictiveness rising).
The Fed raised the target range 25 bps to 3.75%–4.00% on September 16, 2026, a clear shift toward restraint and a signal that policy risk is back on the table. (federalreserve.gov) The implementation details (including a higher rate paid on reserve balances effective September 17) reinforce that this was not symbolic; it was a real tightening impulse likely to show up in credit and interest-sensitive sectors with a lag. (federalreserve.gov)

2) “Hard trigger” recession alarms remain quiet: claims are near cycle lows.
Initial jobless claims fell to 196,000 for the week ending September 12, 2026, the lowest since mid-July and well below levels that typically precede recessionary labor dynamics. (apnews.com) Continuing claims also fell (a useful cross-check that separations aren’t turning into persistent unemployment), consistent with the Sahm Rule staying untriggered (your reading -0.07).

3) Credit is not pricing recession: HY spreads remain tight (financial stress low).
With HY OAS ~270 bps and your Chicago Fed NFCI at -0.56 (loose), financial conditions are not behaving like a pre-recession tape. This is a critical offset to the deterioration in soft/leading indicators: when recessions arrive quickly, they usually do so with a more obvious “price of risk” repricing.

4) Leading labor signal flashing: Temporary Help is deep DANGER.
Temporary help employment (2,520k) remains one of the most recession-sensitive labor series. Employers typically cut temp labor before they cut core payrolls; that makes this a high-signal “early warning” even when claims are still calm. In this model, temp help is doing the heavy lifting on the “caution” case.

5) Household fragility: confidence and the savings cushion are weak.
The University of Michigan tables show a 55.2 reading (September 11, 2026 tables), consistent with a consumer that is still spending but increasingly uneasy. (data.sca.isr.umich.edu) Meanwhile, the personal savings rate at 3.0% (WARNING) implies limited buffer if rates stay high and job growth slows. This matters because a “no recession” path still requires the consumer to absorb higher debt service without a meaningful labor-market break.

6) Markets are calm at the index level, but risk is showing up as valuation stress (and complacency).
The VIX at 17.8 (SAFE) and equity indexes near highs are not recession-confirming. In fact, the post-hike bounce in equities suggests markets may be leaning toward a “growth holds + inflation sticky” equilibrium. The risk is second-order: overvaluation (NASDAQ/GDP in DANGER, S&P 500/GDP in WARNING) increases sensitivity to any earnings disappointment or credit widening.

Category Breakdown

Using your CATEGORY BREAKDOWN counts:

• Primary Indicators: 3 safe / 4 watch / 2 danger
  Mixed, but tilting cautious: “hard” labor/real-time triggers are mostly okay, while leading labor/activity proxies (like temp help) keep risk elevated.

• Secondary Indicators: 2 safe / 0 watch / 1 danger
  Mostly stable; the danger signal here reinforces that the slowdown is broadening beyond one-off noise.

• Housing & Construction: 0 safe / 1 watch / 1 danger
  Housing remains the cleanest interest-rate transmission channel; starts/permitting are consistent with restrictive policy biting at the margin.

• Business Activity: 2 safe / 1 watch / 0 danger
  Business activity isn’t collapsing; this supports the “slowdown, not recession” base case near-term.

• Consumer Credit Stress: 1 safe / 2 watch / 1 danger
  The consumer isn’t breaking yet, but delinquencies + low savings cushion increase downside convexity if the labor market softens.

• Market Signals: 7 safe / 2 watch / 5 danger
  This is the unusual split: markets are calm on volatility/spreads, but flashing “danger” on valuation and macro-asset ratios—classic late-cycle fragility.

• Liquidity: 0 safe / 1 watch / 2 danger
  Liquidity is less forgiving: the depletion of the ON RRP facility and broader liquidity signals imply less shock absorption capacity if stress rises.

• Real-Time / High-Frequency: 0 safe / 1 watch / 1 danger
  High-frequency data is not uniformly bad, but the danger signal indicates the goods economy (freight) is a real soft spot.

Biggest Movers

Top 5 by absolute 7-day % change (from your BIGGEST MOVERS):

1. NY Fed Recession Probability (3.4%): -40.6% (7D)
   Contradictory (improving). The model-implied probability is moving lower, consistent with tight spreads and calm volatility.

2. Freight Transportation Index (-0.3): -40.0% (7D)
   Confirmatory (worsening risk). Freight weakness is consistent with goods-side slowing and often leads broader manufacturing softness.

3. Sahm Rule (-0.07): -30.0% (7D)
   Contradictory (improving). A more negative (safer) reading indicates the unemployment rate isn’t accelerating higher relative to its recent low.

4. Yield Curve (2s10s) (0.27): +25.0% (7D)
   Ambiguous, but mildly confirmatory for late-cycle risk. A steepening after inversion can occur because long rates rise (inflation term premium / fiscal) or because short rates fall (easing). Here, steepening doesn’t automatically mean “recession soon,” but it’s a regime change worth monitoring.

5. VIX (17.8): -12.4% (7D)
   Contradictory (improving). Vol compression signals complacency; it reduces near-term stress but can increase vulnerability to shocks.

90-Day Indicator Trends

No full 90-day window is available in the data you provided; the “90-day history” shown spans roughly June 20 → July 9, 2026 for many series (about 20 days). No data available for this window (full 90 days). Below is a trend read using the available history plus today’s snapshot, focusing on direction-of-travel:

Labor market (hard data): still strong; leading labor: deteriorating.

• Initial claims improved from 226k (Jun 20) to 215k (Jul 9) in your history, and are 196k today—a material improvement and strongly inconsistent with imminent recession.
• Sahm Rule moved from 0.10 (Jun 20) to 0.07 (Jul 9) and is -0.07 today—still firmly in SAFE territory and moving safer.
• Temp help stayed deep DANGER (around 2,490k → ~2,499k in early July history) and is 2,520k today but still flagged DANGER in your framework, implying the level and/or trend relative to benchmark remains recession-consistent even if the last few prints moved.

Financial conditions: loose; valuations stretched.

• NFCI remained around -0.50 to -0.52 in late June/early July and is -0.56 today: conditions look even looser.
• HY OAS oscillated 263 → 283 bps across late June/early July and is ~270 bps today: spreads remain tight and calm.
• Equity proxies were elevated throughout: S&P 500/GDP ~0.235 (warning) was steady; the bigger message is not trend but level risk.
• VIX drifted down in early July (high teens to mid-teens) and is 17.8 today, consistent with the “calm tape” narrative.

Housing: soft.

• Housing starts sat around 1,177k (warning) in your early window; today you have 1,239k (warning)—still weak vs trend even if marginally higher.
• Building permits around 1,410k (watch) in early window; today 1,433k (watch)—again, not collapsing, but not strong enough to offset high rates.

Consumer buffer: low and drifting higher only slightly.

• Savings rate improved from 2.6% (danger) to 3.0% (warning) in late June; today 3.0%—still very low, meaning the consumer is more exposed to any labor-market cooling.

Bottom line on “trend”: hard recession triggers are not deteriorating; the score increase is being driven by policy tightening plus late-cycle fragility (temp help, freight, low savings, weak confidence, valuation risk).

Stock Screener Signals

Today’s quant screen is dominated by “value dividend” flags (ARCC, AIG, BBY, FNF, HMC, T, LTM, BCE) with a smaller set of “oversold growth” (CHTR, TLK). Two messages stand out:

1) Market positioning is leaning defensive and cash-flow oriented.
A cluster of low P/E, dividend-oriented names often shows up when investors want carry + durability rather than high-duration growth. That’s consistent with a macro regime where the Fed is tightening and the market wants income and valuation support rather than multiple expansion.

2) The “oversold growth” pocket signals selective stress, not broad capitulation.
CHTR (RSI 28) and TLK (RSI 30) imply there are idiosyncratic drawdowns inside the market even as index levels remain firm. That kind of internal dispersion fits today’s macro: recession is not priced broadly (spreads/VIX calm), but pockets tied to rate sensitivity and balance-sheet leverage can still break first.

One data quality note for your pipeline: the listed yields (e.g., ARCC 1002%) look like unit/decimal scaling issues rather than investable yields. The directional interpretation (dividend/value clustering) is still useful even if the yield field needs normalization.

Latest Economic Developments

Fed: first hike in years; policy is now a live recession-risk variable again.
The Federal Reserve raised rates on September 16, 2026 to 3.75%–4.00%, explicitly re-tightening financial conditions. (federalreserve.gov) This matters for recession risk less because of the immediate mechanical impact, and more because it increases the probability of a policy overshoot into already-soft confidence and rate-sensitive sectors.

Labor market: jobless claims confirm resilience.
Weekly initial claims fell to 196,000 (week ending September 12), reinforcing the view that layoffs remain rare and recession isn’t imminent on labor-market mechanics. (apnews.com) The claims drop is especially important because it undercuts the near-term “hard landing now” narrative even as leading signals (temp help, freight) weaken.

Markets: post-hike bounce and falling long yields (near-term stress still contained).
On September 17, 2026, the S&P 500 rose about 1.1% to ~7,638, with the 10-year yield noted lower (around 4.93% in that session), implying markets treated the hike as manageable rather than recessionary. (apnews.com) That combination—rate hike followed by equity strength—tends to correspond to “growth holds / inflation sticky,” not “recession imminent,” but it can also reflect complacency that breaks if earnings weaken or credit widens.

Consumer sentiment: weak levels persist.
University of Michigan tables show sentiment at 55.2 (September 11 tables). (data.sca.isr.umich.edu) Even if month-to-month moves bounce, the level is low enough that consumers may be more reactive to shocks (gas/food, job news, credit tightening).

Near-Term Outlook (Next 30 Days)

Base case for the next month (through October 18, 2026): slowdown risk rises modestly, but recession risk remains MODERATE unless labor or credit breaks.

What likely drives the next score move:

• Claims trend: the single cleanest near-term trigger. If initial claims begin printing consistently above ~230k–250k (directionally), the “hard data” story changes quickly. For now, 196k is a strong anchor.
• Credit repricing: watch HY OAS. A move from ~270 bps toward ~350–400 bps would be a meaningful stress signal and would likely push the score higher even without a claims spike.
• Housing prints: starts/permits can deteriorate quickly after a policy tightening impulse. Another weak housing report would reinforce the rate-transmission narrative.
• Fed communications: after the September 16 hike, any guidance that explicitly keeps another hike “live” would raise policy-error risk.

Long-Term Outlook (3-6 Months)

Over the next 3–6 months (through March 2027), the distribution is widening: the most plausible path remains slowdown without recession, but the left tail is fattening because tightening is occurring when buffers are thin (low savings) and leading labor indicators (temp help) are already flashing.

Three structural themes matter:

• Late-cycle labor asymmetry: claims can stay low for months—until they don’t. Temp help deterioration is a classic “early” warning; if it starts to bleed into unemployment and continuing claims, recession odds can rise fast.
• Valuation + liquidity sensitivity: with valuation measures elevated and liquidity signals less supportive (ON RRP depleted), the economy/market may be less able to absorb shocks without spillovers into credit.
• Fiscal and term premium pressure: your fiscal indicators (debt, interest expense) are not near-term recession triggers by themselves, but they can keep long rates higher for longer, which tightens financial conditions even if the Fed pauses.

What to Watch

Specific catalysts and thresholds that can move the score meaningfully:

• Weekly initial claims: watch for a trend reversal (multiple weeks rising) rather than a single print. Today’s anchor is 196k (week ending Sep 12). (apnews.com)
• Sahm Rule: current -0.07 is safe; the key is whether unemployment moves materially above 4.1% and pushes the rule toward trigger territory.
• HY OAS: monitor for a widening regime shift (a durable move above ~325 bps would be the first “yellow flag,” and above ~400 bps would be a serious risk-off signal).
• Housing: starts and permits—if they break lower from already-weak levels, the rate-hike impulse is transmitting strongly.
• Confidence: sentiment is weak at 55.2 in the Sep 11 tables; further erosion would raise the probability of a consumption downshift. (data.sca.isr.umich.edu)
• Market internals: dispersion (oversold pockets) versus index strength—if the tape stays up but credit weakens, that divergence usually resolves bearishly.

Sources

No data available for this window.