5 Charts That Caught the Last 3 Recessions — And What They're Saying in 2026

Financial media loves to debate whether a recession is coming. Most of the debate is noise.

Here's what isn't noise: five indicators that, when they all point in the same direction, have preceded every US recession since 1970 with near-perfect accuracy. Not one of them alone — but the convergence of all five.

We source each one from live FRED data and show you exactly what the numbers are today.

Chart 1: The Yield Curve (10Y-3M Spread)

What it is: The difference between what the government pays to borrow for 10 years vs. 3 months. When short-term rates exceed long-term rates, the curve is "inverted."

Why it matters: An inverted yield curve means bond markets believe the Fed will have to cut rates in the future — because the economy will weaken. Every US recession since 1970 was preceded by an inversion.

Track record: 7/7 recessions preceded by inversion. False positives: 1966 (inversion, no recession for 3 years), 1998 (brief inversion during LTCM crisis).

Where we are now: The curve re-steepened to +0.76% in May 2026 after 26+ months inverted. This is historically a late-cycle signal, not an all-clear.

What to watch: A widening above +1.50% while short-term rates are falling would indicate the market is pricing aggressive rate cuts — the most dangerous late-cycle configuration.

Chart 2: Initial Jobless Claims (4-Week Moving Average)

What it is: The number of people filing for unemployment benefits for the first time, averaged over four weeks to smooth weekly noise.

Why it matters: Jobless claims are the fastest-moving labor market indicator. They spike before unemployment rises, because layoffs happen before the official unemployment rate adjusts. A sustained rise above 300,000 has preceded every post-war recession.

Track record: 11/11 recessions preceded by a sustained rise above 300,000. Average lead time: 3–4 months.

Where we are now: 215,000 — well below the 300,000 warning threshold. The labor market is cooling but not cracking.

What to watch: Three consecutive weeks above 280,000 would be the early warning. A jump to 350,000+ would be the confirmation.

Chart 3: ISM Manufacturing PMI

What it is: A survey of purchasing managers at manufacturing firms. Readings above 50 indicate expansion; below 50, contraction.

Why it matters: Manufacturing turns before the broader economy. ISM below 48 for three consecutive months has preceded every post-WWII recession.

Track record: 11/11 recessions preceded by sub-48 readings. Average lead time: 2–6 months. False positives: several brief dips below 48 that didn't lead to recession (most recently 2015–16). The duration and depth matter.

Where we are now: Hovering near 49 — on the contraction/expansion border. Not yet at recession-warning levels, but deteriorating.

What to watch: A sustained move below 48, especially if accompanied by declining new orders and rising inventories (the classic recessionary configuration in the ISM sub-indices).

Chart 4: BAA Corporate Bond Spread

What it is: The difference between the yield on BAA-rated corporate bonds and 10-year Treasuries. This measures how much extra yield investors demand to hold risky corporate debt vs. safe government bonds.

Why it matters: Credit markets anticipate economic trouble months before GDP data confirms it. When BAA spreads widen beyond 300 basis points (3.00%), lenders are pricing in a significant increase in default risk.

Track record: Spreads above 300 bps preceded 6/7 recessions since 1970. The 2020 recession was preceded by a spike from 200 to 370 bps in 3 weeks — the fastest credit tightening on record.

Where we are now: Tight spreads — corporate credit is calm. No stress signal.

What to watch: A sustained widening above 250 bps would be the early warning. Above 350 bps is a strong recession signal. This indicator can move fast — in 2008, spreads went from 200 to 600+ bps in four months.

Chart 5: The Sahm Rule (Real-Time Recession Indicator)

What it is: Developed by Fed economist Claudia Sahm. When the 3-month moving average of the unemployment rate rises 0.50 percentage points above its 12-month low, the economy is in recession.

Why it matters: Unlike the yield curve (which predicts) or ISM (which leads), the Sahm Rule is a real-time detector. By the time it triggers, the recession is usually already underway. But it has never triggered outside of a recession.

Track record: 0 false positives since 1970. Triggered during every recession, typically 0–3 months after the official start date. An incredibly reliable confirmation signal.

Where we are now: The current Sahm Rule reading is 0.13 — well below the 0.50 trigger. The unemployment rate's 12-month low is 4.1%; the current 3-month average is 4.23%. No trigger.

What to watch: If unemployment rises to 4.6%+ and stays there for 2–3 months, the Sahm Rule will trigger. At that point, recession is almost certainly underway.

The Convergence Signal

No single indicator is perfect. But when all five point in the same direction, the signal is extremely reliable.

Current status (May 2026):
- ✅ Yield Curve: Warning (re-steepening phase)
- ✅ Jobless Claims: All-clear
- ⚠️ ISM Manufacturing: Borderline
- ✅ Credit Spreads: All-clear
- ✅ Sahm Rule: All-clear

Net assessment: Elevated caution, not alarm. Two of five indicators are flagging. Historically, a recession signal requires four or five to converge. We're not there yet — but the manufacturing softness and yield curve dynamics warrant close monitoring.

The recession.today composite score, which weights all five indicators plus 16 others, currently reads 28/100 (Low Risk). Want to understand how the score is calculated? Read our methodology deep-dive.