Convex Recession Probability Index (CRPI)
A composite leading indicator that synthesizes five proven recession signals into a single 0-100 reading.
Current Reading
Some indicators deteriorating. Typically reflects an isolated signal (e.g., yield curve flattening) without broader confirmation.
Last updated: Apr 5, 2026
Historical Chart
Methodology
The CRPI was designed to solve a real problem: professional economists track dozens of recession indicators individually, but no single metric captures the full picture. The yield curve can invert for two years before a recession arrives. The Sahm Rule only triggers once unemployment is already rising. Credit spreads can stay tight until the last moment, then blow out overnight.
CRPI synthesizes five distinct recession signals — each measuring a different transmission channel — into one composite reading. By weighting them equally and normalizing to a 0-100 scale, it captures deterioration across multiple channels simultaneously. A reading of 60+ means the majority of recession transmission mechanisms are flashing warning, not just one isolated indicator.
The index requires at least three of five components to compute, making it resilient to individual data gaps or discontinued series.
Components
Uses the worst of the 10Y-2Y and 10Y-3M Treasury spreads. Scores rise as the curve inverts deeper.
The yield curve has inverted before every US recession since 1970. It captures credit channel stress — when short rates exceed long rates, bank profitability collapses and lending contracts.
Measures how close the unemployment rate acceleration is to the 0.5pp Sahm threshold. Scores rise as unemployment accelerates.
Created by Fed economist Claudia Sahm, this indicator has a perfect track record of identifying recession starts in real time. It captures the labor market deterioration channel.
Measures the 3-month percentage change in the 4-week moving average of initial unemployment claims.
Initial claims are the highest-frequency labor market indicator available (weekly). Rising claims are among the earliest signs of economic stress, often leading payroll data by 2-3 months.
Uses the z-score of high-yield credit spreads vs. their 1-year history. Scores rise as spreads widen relative to recent norms.
Credit markets are forward-looking — HY spreads reflect real-time market pricing of default risk. Using a z-score instead of absolute levels avoids bias from secular spread compression.
Momentum of the Conference Board Leading Economic Index. This component was discontinued in February 2020.
The LEI aggregated 10 leading indicators (building permits, stock prices, credit conditions, etc.). Post-discontinuation, CRPI operates on 4 components with adjusted normalization.
Formula
Each component scores 0-20 points. Raw scores are summed and divided by the maximum possible score (number of available components times 20), then scaled to 0-100. This normalization means the index functions correctly whether 3, 4, or 5 components are available.
How to Read CRPI
Expansion intact. No recession signals present across any measured channel.
Some indicators deteriorating. Typically reflects an isolated signal (e.g., yield curve flattening) without broader confirmation.
Multiple warning signals active. At least 2-3 recession channels showing stress. Warrants defensive positioning.
Most indicators flashing red. Historically associated with late-cycle deterioration or active economic contraction.
Near-certain recession conditions. All available channels signaling simultaneously.
What Makes This Different
- -Unlike single-metric recession indicators (yield curve, Sahm Rule), CRPI requires confirmation across multiple channels before reaching elevated levels.
- -The z-score normalization on credit spreads avoids the secular compression bias that makes absolute spread levels misleading over long time horizons.
- -Equal-weight component design means no single indicator can dominate the composite — a 60+ reading truly means broad-based deterioration.
Historical Performance
2018 Q4 Selloff
September 2018 - January 2019The S&P 500 fell 20% as trade war escalation combined with aggressive Fed rate hikes. CRPI moved from Low (12) to Elevated (45) during the selloff — capturing the yield curve flattening and HY spread widening. Crucially, the Sahm Rule and initial claims components stayed low (no labor market deterioration), which is why CRPI peaked at 45 rather than higher. This correctly signaled a growth scare, not a recession.
| Date | Moment | CRPI |
|---|---|---|
| 2018-09-01 | 1 month before peak | 20Moderate |
| 2018-10-01 | Market peak | 12Low |
| 2018-12-24 | Market trough | 45Elevated |
| 2019-01-31 | Recovery begins | 41Elevated |
COVID Crash
January 2020 - June 2020The fastest bear market in history saw the S&P 500 fall 34% in 23 trading days. CRPI was at just 15 (Low) when the market peaked on February 19 — the index could not have predicted an exogenous pandemic shock. But it tracked the damage accurately: rising to 39 at the first circuit breaker (March 9), 55 at the trough (March 23), and peaking at 74 (High) by June as the Sahm Rule triggered and claims data caught up. This illustrates a key limitation: during sudden shocks, CRPI lags because labor data is weekly/monthly, not real-time.
| Date | Moment | CRPI |
|---|---|---|
| 2020-01-15 | Pre-crash baseline | 20Moderate |
| 2020-02-19 | All-time high | 15Low |
| 2020-03-23 | Market trough | 55Elevated |
| 2020-06-01 | CRPI peak (74) | 74High |
2022 Bear Market
January 2022 - January 2023The Fed tightened monetary policy aggressively, raising rates from 0% to 4.5% and beginning quantitative tightening. The S&P 500 fell 25% over nine months. CRPI rose from 6 (Low) at the January peak to 42-45 (Elevated) by March — correctly capturing the yield curve inversion and credit spread widening. It stayed at Elevated (42-45) throughout the bear market. Notably, it never reached High because the Sahm Rule didn't trigger (unemployment remained low even as markets crashed), correctly distinguishing this as a valuation reset, not a recession.
| Date | Moment | CRPI |
|---|---|---|
| 2022-01-03 | Market peak | 6Low |
| 2022-03-15 | First rate hike | 42Elevated |
| 2022-10-12 | Market trough | 45Elevated |
| 2023-01-31 | Recovery underway | 45Elevated |
2024 Carry Trade Unwind
July 2024 - September 2024The JPY carry trade unwind triggered a global equity selloff, with VIX spiking to 65. CRPI was already at 66-67 (High) before the event — the deeply inverted yield curve and Sahm Rule triggering at 0.57 had pushed it to elevated levels. It peaked at 82 (Very High, its all-time maximum) on August 5. This was CRPI's strongest performance: it was signaling High risk weeks before the selloff materialized, providing genuine advance warning.
| Date | Moment | CRPI |
|---|---|---|
| 2024-07-01 | CRPI already at High (67) | 67High |
| 2024-07-16 | S&P 500 peak | 66High |
| 2024-08-05 | CRPI all-time high (82) | 82Very High |
| 2024-09-30 | Market recovery | 54Elevated |
Limitations
- -The Sahm Rule and initial claims components are coincident or slightly lagging indicators — they respond to layoffs already happening, not ones about to happen.
- -The Leading Economic Index component was discontinued in February 2020, reducing the index to 4 components for the post-2020 period.
- -During sudden exogenous shocks (like COVID), CRPI may spike after the market has already bottomed, because labor market data takes weeks to reflect the damage.
- -Yield curve inversion can persist for years before a recession materializes, contributing to sustained mid-range readings during the lag period.
Related Metrics
Get CRPI updates in your inbox
Daily indicator readings with macro context. Free.
This indicator is generated from live economic data and is for informational purposes only. It does not constitute financial advice. Past performance does not guarantee future results.