Summary¶
Over roughly five weeks (February 19 – March 23, 2020), the S&P 500 fell about 34%, the VIX reached its then-highest close ever (~82 on March 16), and trading halted on Cboe on March 18 under a Level 1 (7%) market-wide circuit breaker. The event was a textbook volatility regime shift: implied volatility, correlations, and margin requirements all repriced together, and liquidity in both equities and options thinned exactly when hedging demand peaked.
Timeline¶
| Date | Event |
|---|---|
| Feb 19, 2020 | S&P 500 record high; VIX ~14 — historically calm regime |
| Feb 24–28 | COVID goes global; S&P falls ~12% in a week; VIX doubles into the 40s |
| Mar 9 | Oil price war + COVID; limit-down session; VIX > 54 |
| Mar 12, 16 | Repeated limit-downs; VIX closes ~82.7 (Mar 16) — record |
| Mar 13–18 | Fed emergency cuts, QE expansion; OCC and CME raise margin requirements repeatedly |
| Mar 18 | Level 1 (7%) market-wide circuit breaker halts trading at ~3:25 pm ET |
| Mar 23 | Bottom of the decline; recovery begins |
Mechanics¶
- Vol regime shift: a quiet, carry-harvesting regime ended abruptly. Vol-of-vol exploded; historical correlations converged toward 1 across asset classes — the diversification failure described in Diversification and Correlation at maximum amplitude.
- Margin spiral: clearinghouses (OCC/CME) raised requirements after the vol spike, forcing de-risking into falling, illiquid markets. Margin is a liquidity risk independent of max loss (see VaR and Margin).
- Liquidity evaporation: bid-ask spreads in options widened sharply; many short-option marks became model-based rather than executable. Stops and delta-replication hedges executed into gaps — the Natenberg discrete-hedging failure in live conditions (see Tail Risk Principles).
- EWMA/GARCH updating: trailing-500-day VaR inputs lagged badly; vol- and correlation-updated models captured the regime change far earlier — Hull's September 2008 lesson repeating.[^hull-var]
Who Got Hurt and Why¶
Illustrative outcomes (hypothetical, not empirical). The rows below are qualitative illustrations of loss mechanisms, not established empirical statistics; the cited references do not quantify trader outcomes in this event.
| Group | Outcome | Why |
|---|---|---|
| Naked short put / condor sellers | Account failures were a live risk | Uncapped or wing-wide losses × IV doubling × margin hikes |
| Defined-risk condor sellers (sized to 2%) | Drawdowns contained by design | Max loss known; heat-limited; early-exit rules ([^tomic-risk]) applied rather than riding to the cap |
| Long-vol / "units" holders | Potential convex gains | Cheap OTM puts and VIX calls snowballed far beyond model value[^tomic-risk] |
| Cash-heavy traders | Dry powder as a hedge | Cash as a position: ability to skip marginal trades and redeploy into the recovery |
| Levered inverse-Vol ETP traders (post-2018 rules) | Mixed (structurally dependent) | Volmageddon-era termination clauses and structural changes changed the loss profile |
Lessons¶
- Regime shifts are discontinuous: a strategy calibrated to the prior regime (low IV, low correlation) fails without regime detection — the motivation for Regime-Dependent Delta Exposure and the regime concepts in 60-regimes.
- Margin is the tail's delivery mechanism: positions that could be held to recovery were liquidated by raised requirements; liquidity buffers and sizing headroom are tail defenses, not inefficiencies.
- Defined-risk sizing works as designed (illustrative): the pattern above — defined-risk accounts honoring TOMIC's per-trade and heat limits holding through the event while margin-sized accounts faced forced liquidation — is an illustrative strategy expectation consistent with the case's mechanics (capped losses, heat limits, margin as liquidity risk), not an empirical survivor statistic.[^tomic-risk]
- Exit early, not at the cap: the third-third-third rule converts a tail event into a controlled, smaller loss when acted on in the first third.[^tomic-risk]
- Update risk inputs, don't just reprice: EWMA/GARCH-class vol and correlation updating is the difference between a VaR that flagged the regime and one that didn't.[^hull-var]
References¶
- Federal Reserve, FOMC statements and emergency actions, March 3/15/23, 2020, federalreserve.gov.
- OCC, notices of margin requirement increases, February–March 2020, occ.org.
- Cboe Global Markets, market-wide circuit breaker halt notice, March 18, 2020, cboe.com.
- CME Group, margin requirement adjustment notices, March 2020, cmegroup.com.
- Baker, S., Bloom, N., Davis, S., et al., "The Unprecedented Stock Market Reaction to COVID-19," Review of Asset Pricing Studies, 10(4), 2020 — documented magnitude and speed of the drawdown and vol spike.
Links¶
- 60-regimes index — regime framework
- Regime-Dependent Delta Exposure
- Tail Risk Principles — defined-risk sizing and unit hedges
- VaR and Margin — margin as liquidity risk, vol-updated VaR
- 2018 Volmageddon, August 2024 Spike