Macro Strategy · 8 min read · 2026-05-01
The 9 leading indicators that front-run a recession.
No single indicator predicts recessions reliably. Nine, scored as a composite, do.
Every recession is obvious in retrospect.
The S&P 500 has lost a median 32% in the bear market that accompanies a U.S. recession since 1970. Every recession was preceded by clear leading indicators. Reading them in real time, before they have all confirmed, is the harder problem.
The nine indicators
The nine leading indicators of a U.S. recession.
Each has its own lead time and its own false-positive rate. Composite scoring across all nine produces a more reliable signal than any single indicator.
10y minus 3m yield curve inverted for 90+ days
Threshold: spread < 0 for 90d
The most reliable single recession predictor since 1970. Lead time 6–18 months. Two false positives (1966, 2022 — TBD).
Sahm Rule trigger: U3 unemployment rate up 0.5% from trailing 12-month low
Threshold: triggered
Real-time recession indicator developed by Claudia Sahm. Has identified every U.S. recession since 1970 within 1–4 months of onset, with no false positives.
ISM Manufacturing PMI below 47 for 3+ months
Threshold: < 47, persistent
Sustained sub-47 readings have preceded most recessions in the modern record. PMI alone is noisy; a three-month confirmation reduces false positives sharply.
Initial jobless claims 4-week MA up 25% from cycle low
Threshold: +25% from low
Claims rise before payrolls fall. A 25% increase from the cycle low historically precedes recessions by 4–10 months.
High-yield credit spreads above 500 bps and widening
Threshold: HY OAS > 500 bps
Credit markets price recession risk before equities. HY spreads above 500 bps signal stressed funding conditions; widening from there is the credit-cycle turn.
LEI (Conference Board Leading Economic Index) down 6 of last 12 months
Threshold: 6/12 negative
Composite of 10 leading indicators. A six-of-twelve decline has preceded every recession since the 1960s with a 6–12 month lead time.
Conference Board CEO Confidence Index below 35
Threshold: < 35
CEO sentiment translates directly to capex and hiring decisions. Below 35, the cohort is preparing for contraction.
M2 money supply contracting year-over-year
Threshold: YoY < 0
Annual M2 contractions are rare and have been associated with major economic transitions. Combined with the other indicators, the contraction adds confidence.
Cass Freight Index down 6%+ year-over-year
Threshold: YoY < -6%
Freight is a real-time consumption indicator. A 6%+ annual decline persists for 3+ months in only the worst environments.
Why composite scoring beats any single indicator
Each leading indicator has its own track record, its own lead time, and its own false-positive rate. The yield curve has a perfect record of preceding every postwar U.S. recession but has produced two arguable false positives. The Sahm rule has identified every recession at onset but offers no advance warning. ISM PMI is noisy month-to-month and produces frequent false alarms below 50. Each indicator is informative; none is sufficient alone.
A composite score across the nine resolves the trade-offs. When seven or more of the nine fire, the conditional probability of a recession within 12 months has historically exceeded 75 percent. When three or fewer fire, the conditional probability drops below 10 percent. The middle band — four to six firing — is the regime of greatest uncertainty and the regime in which positioning matters most.
The yield curve, properly read
The 10-year minus 3-month spread is the most reliable recession predictor in the historical record. The signal is the persistence of inversion, not the depth. A brief intraday inversion is noise. A sustained inversion of 90 days or more has preceded every postwar U.S. recession with lead times between 6 and 18 months.
The 2022 inversion is the live test of the indicator's continuing validity. The traditional read suggests recession; the post-COVID environment has produced a series of countervailing factors that may have delayed or muted the response. The composite framework allows the yield curve to be one of nine inputs rather than the single decision.
The Sahm rule — real-time, not leading
Claudia Sahm's rule fires when the three-month-average unemployment rate rises 0.5 percentage points or more above its trailing 12-month low. The rule does not lead recessions; it identifies them in real time, typically within one to four months of NBER's eventual dating. The value is that it has no historical false positives.
In the composite framework, the Sahm rule serves as the confirmation indicator. When the leading indicators have been firing for months and the Sahm rule then triggers, the composite has reached its highest-confidence configuration. Equity drawdowns historically begin shortly before or shortly after this point.
Credit spreads — the bond market's leading signal
High-yield credit spreads tend to widen before equities decline. The mechanism is funding-driven: when leveraged corporate borrowers face rising spreads, refinancing becomes more difficult, capex is cut, and the slowdown begins to register in macro data months later. Spreads above 500 basis points and widening are the threshold most consistent with the start of a credit cycle turn.
The 2008 cycle widened from 300 bps to over 2,000 bps in roughly 14 months. The 2020 cycle widened to roughly 1,100 bps before snapping back. The 2024–2025 environment kept spreads compressed in the 250–350 bp range — the absence of widening was itself an informative non-signal.
Positioning around the composite score
The composite framework is a positioning tool, not a market-timing tool. The recession is not a date; it is a regime. When the composite scores 7+, the appropriate response is defensive: increase bond ladder length, reduce equity beta, raise cash, and avoid initiating new high-multiple positions. When the composite scores 3 or fewer, normal equity exposure is appropriate.
The composite is updated monthly as the underlying data releases. Most of the indicators have monthly or weekly updates, with some (PMI, claims) updating more frequently. A disciplined monthly review takes 30 minutes and produces a composite score that has historically correlated with forward 12-month equity returns at roughly -0.55.
What this framework does not predict
The composite predicts the U.S. business cycle. It does not predict equity market tops or bottoms with timing precision. The S&P 500 has historically peaked 3 to 9 months before NBER recession onset and bottomed 3 to 6 months after the recession peaks in macro data. The framework gives a regime read; the entry and exit timing of equity positions requires additional tools.
Get the nine, every Monday.
Free weekly digest. The only U.S. equity letter that publishes a name only when nine independent signals align.
Common questions
Questions.
Has the yield curve given any false positives?
Two arguable false positives in the modern record: a brief inversion in 1966 not followed by recession in the immediate window, and the 2022 inversion which has not yet been definitively followed by NBER recession dating.
Does the Sahm rule always fire at recession onset?
Every U.S. recession since 1970 has triggered the Sahm rule within 1–4 months of NBER onset. No false positives in the historical record.
How do I track the composite without spending all day on Bloomberg?
FRED (St. Louis Fed) publishes most of the inputs for free. A monthly review of FRED's recession dashboard plus the Conference Board LEI release covers eight of the nine. Cass Freight Index publishes monthly via press release.
What about international recession indicators?
Versions of the framework exist for European, Japanese, and Chinese cycles. The yield curve signal is universal; ISM-equivalent surveys (ZEW, Tankan, Caixin) substitute for ISM PMI. The lead times tend to be similar.
Should I exit equities entirely when the composite scores 7+?
Not necessarily. Equity drawdowns into recessions average 30%+ but vary widely. Reducing exposure, lengthening bond ladders, and avoiding leveraged or high-beta names captures most of the defensive benefit without the timing risk of full exit.
What's the false-positive rate of the full composite?
Approximately 8% in the post-1970 sample — significantly lower than any single indicator. The composite trades a slightly later signal for a much cleaner one.
One name. Sometimes weeks of silence. Always with conviction.
Free weekly digest. No credit card. Unsubscribe anytime.