Convex Risk Appetite Index (CRAI)
A cross-asset risk appetite measure built from price ratios, what markets are actually doing, not what surveys say.
Current Reading
Balanced risk-taking. No strong directional signal from cross-asset flows.
Last updated: May 18, 2026
Historical Chart
Methodology
Most sentiment indicators rely on surveys (AAII, NAAIM) or single-asset volatility (VIX). These are useful but limited, surveys lag actual positioning, and VIX only captures equity option demand. CRAI takes a fundamentally different approach: it measures what investors are actually buying and selling across five distinct risk-on/risk-off pairs.
The insight is simple: when investors are truly risk-seeking, money flows simultaneously into small caps over large caps, high-yield bonds over investment-grade, consumer discretionary over staples, emerging markets over developed, and regional banks over the broad market. When any one of these pairs diverges, it reveals selective risk-taking. When all five align, it confirms genuine broad-based appetite (or aversion).
Each pair is expressed as a price ratio with a 60-day z-score. This relative approach means CRAI adapts to the current regime, a given IWM/SPY ratio level means different things in a bull vs bear market. The z-score captures whether the ratio is unusually high or low relative to its recent behavior, which is what actually matters for positioning.
Components
60-day z-score of the IWM/SPY price ratio. Higher values mean small caps are outperforming, indicating risk appetite.
Small caps are more sensitive to economic conditions, credit availability, and domestic growth expectations. When investors rotate from large caps to small caps, it signals confidence in the economic cycle.
60-day z-score of the HYG/LQD price ratio. Higher values mean investors prefer risky credit over safe credit.
The credit market is the most direct measure of risk appetite in fixed income. When investors are willing to hold low-rated corporate debt over high-rated debt, they are explicitly pricing in lower default risk.
60-day z-score of the XLY/XLP price ratio. Higher values mean consumers are spending on discretionary goods.
This ratio captures consumer confidence and spending patterns. Discretionary outperformance signals that investors expect strong consumer spending, while staples outperformance signals defensive positioning.
60-day z-score of the EEM/EFA price ratio. Higher values mean capital is flowing toward higher-risk emerging markets.
EM equities carry currency risk, political risk, and higher beta to global growth. Capital flowing to EM over DM signals genuine global risk appetite, not just a US-centric equity rally.
60-day z-score of the KRE/SPY price ratio. Higher values mean regional banks are outperforming.
Regional banks are highly sensitive to yield curve shape, credit conditions, and commercial real estate. Their outperformance signals confidence in the financial system and lending cycle.
Formula
For each pair, compute the 60-day z-score of the price ratio. Map z-scores to 0-20 points: z of -2 yields 0 points (maximum risk aversion), z of +2 yields 20 points (maximum appetite). Sum component scores and normalize to 0-100. Requires at least 3 of 5 components.
How to Read CRAI
Extreme risk aversion across all channels. Investors flee to safety simultaneously. Often a contrarian buy signal.
Defensive positioning dominant. Capital favoring safe havens over risk assets across most pairs.
Balanced risk-taking. No strong directional signal from cross-asset flows.
Broad risk-on positioning. Money flowing into small caps, HY, discretionary, EM, and banks simultaneously.
Extreme risk appetite / complacency. All channels showing risk-on simultaneously. Watch for reversal signals.
What Makes This Different
- -Unlike survey-based sentiment (AAII, NAAIM), CRAI measures what investors are actually doing with their money, not what they say they're doing.
- -Unlike VIX, which only captures equity option demand, CRAI spans five distinct asset classes, equities, credit, sectors, international, and banking.
- -The z-score normalization means CRAI adapts to the current regime. A reading of 70 always means "unusually risk-on relative to the past 60 days" regardless of the absolute market level.
- -CRAI uses completely independent data sources from CVRP and CNLI, ETF prices rather than FRED economic data, making the trifecta of Convex indices truly complementary.
Cross-Index Playbook (live)
Current Regime: Risk-On Expansion
calmFundamentals, positioning, and liquidity are broadly supportive of risk assets.
CVRP at 9 indicates low recession probability. CRAI at 52 shows healthy risk appetite without extremes. CNLI at $5.89T provides ample liquidity. NVI at 88 suggests the narrative environment is accelerating, a regime change may be forming. This is the environment where staying invested and riding trends pays off.
- -Stay invested in risk assets. The macro backdrop is supportive.
- -Focus bottom-up: with no macro headwind, stock and sector selection drives returns.
- -Don't overthink or hedge excessively, the cost of protection drags returns in benign environments.
- -Watch CVRP and NVI for early deterioration signals. Regimes shift fastest when least expected.
CRAI at 52 and CVRP at 9, risk appetite is backed by solid fundamentals. This is the "buy the dip" regime where pullbacks are opportunities, not warnings. Stay constructive until either index flashes a warning.
This regime assessment is generated algorithmically from live Convex index values. It is not financial advice. Cross-index signals add context but do not guarantee outcomes. Always consider your own risk tolerance, time horizon, and the full market picture.
How to Use This Signal
How CRAI Triggers Position Changes
CRAI is most useful when interpreted as a short-cycle (1-3 month) regime indicator with three actionable zones, distinct from CVRP's multi-quarter horizon and CNLI's structural-liquidity timescale. The index reflects what investors are actually doing across five risk-on/risk-off pairs, making it a positioning tool more than a forecast tool.
Zone 1: CRAI 0-30 (Very Low to Below Average). This is the panic zone, and historically the cleanest contrarian buy signal. CRAI dropped to 11 during the SVB banking crisis (March 13, 2023) and to 18 at the 2022 bear-market trough. In both cases, broad-based risk aversion across all five pairs (small caps, HY credit, discretionary, EM, regional banks all simultaneously selling) marked the capitulation low. Action: aggressive risk-on positioning is appropriate, especially in the most-sold pairs. Add to small caps (IWM), buy regional bank exposure (KRE), HY credit (HYG), discretionary (XLY), and EM (EEM). CRAI re-entries above 30 within 4-8 weeks of the trough have historically produced 10-20% short-term gains in the relevant pairs.
Zone 2: CRAI 30-70 (Below Average to Elevated). This is the normal-regime zone, where positioning should match the directional trend rather than fight it. Action: maintain neutral-to-modestly-overweight risk positioning, with overlay tilt toward whichever pair is currently leading (if HYG/LQD is rising while IWM/SPY is flat, lean into credit risk over equity risk).
Zone 3: CRAI 70-100 (Elevated to Very High). This is the complacency zone, and historically a contrarian sell signal as it accumulates. CRAI above 80 sustained for weeks (rare, occurred briefly in 2017 and late 2024) has preceded the 2018 Q4 selloff and the August 2024 carry-trade unwind. Action: tighten stops on existing risk positions, add modest hedges (long volatility through VIX calls, partial collars on equity exposure), reduce small-cap overweight in favor of large-cap quality. The cost of being defensive in this zone is small (most of the gains in the highest-CRAI period have already been captured); the cost of being unhedged when the regime turns is large.
The most actionable signals come from CRAI direction at extremes: rolling up from below 30 (full risk-on signal) or rolling down from above 70 (warning signal).
Setup 1: March 2023 SVB Capitulation Low
In the week ending March 13, 2023, CRAI plunged from 69 (Elevated) on March 1 to 11 (Very Low) by March 13 as Silicon Valley Bank collapsed and triggered systemic banking concerns. All five pairs simultaneously shifted to risk-off: KRE/SPY collapsed (regional banks selling vs broad market), HYG/LQD dropped (credit stress), IWM/SPY fell (small-cap risk aversion), XLY/XLP defensive rotation, EEM/EFA EM weakness. The Z-score breadth of the move was extreme: every component hit -2 standard deviations or worse simultaneously, the textbook capitulation signature.
Action that worked: aggressive contrarian buying in the most-sold areas. KRE bottomed near $36 on March 13 and rallied to $48 by May 1 (+33% in seven weeks). HYG bottomed and rallied 4% over the next month as HY OAS compressed from 520bp to 460bp. CRAI re-entered above 30 by April 1 (signaling regime exit) and above 50 by mid-April. Investors who used the CRAI trough as their re-entry signal captured the cleanest part of the post-SVB rebound. The setup is the canonical example of CRAI's value: it correctly identified breadth (all five pairs moving together = systemic, not idiosyncratic) and timed the re-entry better than VIX (which only reached 26 during the episode, less extreme than CRAI's reading suggested).
Setup 2: 2024 Carry-Trade Unwind (Less-Extreme but Faster)
Between July 16 and August 5, 2024, CRAI fell from 58 (Neutral) to 28 (Below Avg) as the JPY carry-trade unwind hit global risk assets. The decline was sharp but importantly less extreme than SVB, telling investors that the breadth of the selloff was narrower (concentrated in equity-positioning unwinds, not in credit or fundamental risk repricing). HYG/LQD barely flinched; EEM/EFA held up; KRE/SPY actually outperformed during the episode.
Action that worked: partial risk-on response rather than aggressive contrarian buying. Adding to broad equity exposure (SPY) during the August 5 panic worked: S&P 500 rallied +12% from the August 5 low through year-end. But adding aggressively to small caps or regional banks (which CRAI did not signal as oversold) was unnecessary. The cycle illustrates CRAI's utility as a breadth indicator: a moderate trough (28) signals a moderate buying opportunity (broad index rather than aggressive deep-cyclical), while an extreme trough (11 in SVB) signals a deep-value buying opportunity (specific oversold pairs). The differentiation is what survey-based or single-asset sentiment indicators cannot provide.
What CRAI Is Saying Right Now (April 2026)
CRAI in April 2026 is estimated in the 50-65 range (Neutral to Elevated), reflecting balanced positioning across the five pairs after the early-2026 BTC drawdown and partial equity correction. The setup is no longer at extreme complacency (the late-2025 readings approaching 80 have moderated) but is also not signaling stress.
Component reads as of April 2026: IWM/SPY ratio is mid-range (small caps not leading, but not collapsing); HYG/LQD ratio remains tight (credit complacency intact, supports above-50 reading); XLY/XLP is mid-range with the consumer-discretionary trade neither leading nor lagging materially; EEM/EFA has been flat-to-rising (EM resilience contributes positively); KRE/SPY has stabilized after early-2026 weakness (banks not panicking despite mixed Q1 earnings).
The April 2026 reading is consistent with mid-cycle positioning: investors maintaining risk exposure but not adding aggressively, with hedges still trimmed but not fully off. Watch for direction: a move toward 70-80 would warn of complacency-driven topping risk, while a move below 40 would signal the start of a broader risk-off regime that would warrant more aggressive defensive action.
Action: maintain neutral-to-mild risk-on positioning consistent with mid-cycle CRAI. Re-allocation triggers tied to CRAI rolling above 75 (start tightening hedges and trimming high-beta) or below 35 (begin contrarian risk-on adds in oversold sub-pairs). Use CRAI as the timing overlay for cross-asset rebalancing decisions that CVRP and CNLI generate at slower frequencies.
Historical Performance
SVB Banking Crisis
March 2023CRAI produced its most dramatic signal during the SVB collapse. From a reading of 69 (Elevated) on March 1, it plunged to 11 (Very Low) by March 13 as all five pairs simultaneously shifted to risk-off. KRE/SPY collapsed (banks vs broad), HYG/LQD dropped (credit stress), and IWM/SPY fell (small cap risk aversion). This was a faster and more dramatic signal than VIX, which only reached 26. The breadth of the collapse across all five channels correctly identified this as a systemic risk event, not just a bank stock selloff.
| Date | Moment | CRAI |
|---|---|---|
| 2023-03-01 | Pre-crisis (69) | 69Elevated |
| 2023-03-08 | SVB stock crash | 37Below Avg |
| 2023-03-13 | CRAI trough (11) | 11Very Low |
| 2023-05-01 | Recovery (55) | 26Below Avg |
2024 Carry Trade Unwind
July - September 2024The JPY carry trade unwind saw VIX spike to 65, but CRAI told a more nuanced story. It dropped from 58 (Neutral) to 28 (Below Avg), a significant move, but not the extreme seen during SVB. Why? Because HYG/LQD barely flinched (credit was fine) and EEM/EFA held up (EM wasn't selling off). CRAI correctly identified this as an equity-concentrated event rather than broad-based risk aversion, and the market recovered within weeks.
| Date | Moment | CRAI |
|---|---|---|
| 2024-07-16 | S&P 500 peak (58) | 83Very High |
| 2024-08-05 | VIX spike day (28) | 36Below Avg |
| 2024-09-30 | Full recovery (61) | 67Elevated |
2022 Bear Market
January - October 2022During the Fed tightening cycle, CRAI provided sustained warning. It dropped from 65 (Elevated) in January to 22 (Below Avg) by June, and stayed depressed throughout the summer. Every risk-on pair was aligned: small caps underperforming, HY selling vs IG, discretionary weak vs staples, EM bleeding out vs DM. The sustained low reading, not a spike, correctly signaled a structural risk-off environment rather than a temporary panic.
| Date | Moment | CRAI |
|---|---|---|
| 2022-01-03 | Market peak (65) | 46Neutral |
| 2022-06-15 | Mid-bear (22) | 64Elevated |
| 2022-10-12 | Market trough (18) | 63Elevated |
| 2023-01-31 | Recovery begins (48) | 65Elevated |
Limitations
- -CRAI relies on ETF price data, which is only available from April 2021 onward, shorter historical coverage than FRED-based indicators.
- -The 60-day z-score lookback means CRAI adapts slowly to regime changes. In the first weeks of a new regime, the z-scores are still anchored to the old baseline.
- -During market dislocations with extreme correlations (2020 COVID), all five pairs can move together, making CRAI swing between extremes rather than providing nuanced readings.
- -ETF price ratios can be distorted by fund flows, rebalancing, and creation/redemption mechanics independent of fundamental risk appetite.
Related Metrics
Explore Further
Frequently Asked Questions
What is the Convex Risk Appetite Index (CRAI)?▾
A cross-asset risk appetite measure built from price ratios, what markets are actually doing, not what surveys say. Full methodology, components, and historical case studies are published on this page for transparency.
What components go into CRAI?▾
CRAI combines 5 components: Small Cap vs Large Cap (IWM/SPY), High Yield vs Investment Grade (HYG/LQD), Discretionary vs Staples (XLY/XLP), Emerging vs Developed Markets (EEM/EFA), and others. Each component is documented with its source series, weight, and the reason it was included in the index. For each pair, compute the 60-day z-score of the price ratio. Map z-scores to 0-20 points: z of -2 yields 0 points (maximum risk aversion), z of +2 yields 20 points (maximum appetite). Sum component scores and normalize to 0-100. Requires at least 3 of 5 components.
How do I interpret a CRAI reading?▾
The index uses a scaled reading with labelled zones: 0-20 = Very Low; 20-40 = Below Avg; 40-60 = Neutral; 60-80 = Elevated; 80-100 = Very High. The current reading, its zone, and the 30-day and 90-day change are displayed at the top of this page. Historical case studies below show how the index has behaved during prior regime transitions.
How often is CRAI updated?▾
CRAI recalculates when its underlying component series publish new data. Most components refresh daily or weekly, so the composite reading is typically no more than one business day stale. The "Last updated" timestamp next to the current reading shows the exact observation date.
What are the known limitations of CRAI?▾
No single composite indicator captures every risk channel. The documented limitations for CRAI are listed in the Limitations section on this page. Convex publishes these explicitly so that readers can form their own view about when the index is most and least informative. For a cross-check against other Convex indices, see the Cross-Index Playbook below.
How does CRAI differ from comparable third-party indicators?▾
Unlike survey-based sentiment (AAII, NAAIM), CRAI measures what investors are actually doing with their money, not what they say they're doing. Unlike VIX, which only captures equity option demand, CRAI spans five distinct asset classes, equities, credit, sectors, international, and banking. The full list of what makes CRAI distinctive is on this page under "What Makes This Different." Convex publishes the formula so any researcher can replicate the calculation, rather than treating the index as a black box.
Get CRAI 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.