Volmageddon: The February 5, 2018 Volatility Regime Break

Executive Summary#

On February 5, 2018, the CBOE Volatility Index (VIX) experienced the largest single-day percentage increase in its history, spiking 115.6% from an intraday low near 17 to a closing value of 37.32, with after-hours indicative prints exceeding 50. This event, colloquially termed “Volmageddon,” annihilated approximately $6 billion in notional value across short-volatility exchange-traded products (ETPs) within hours and caused cascading dislocations in options spread pricing across the S&P 500 term structure.

1. Pre-Event Regime: The Compressed Volatility Equilibrium#

1.1 VIX Structural Suppression (2016–Jan 2018)#

The 2017 calendar year was characterized by historically anomalous realized volatility compression:

This regime created a persistent, elevated variance risk premium (VRP), defined as:

$$ VRP = \sigma^2_{\text{implied}} - \sigma^2_{\text{realized}} $$

With the VIX near 10 and 20-day realized vol near 6%, the annualized VRP stood at roughly 64 variance points (100 − 36), making systematic short-vol strategies extraordinarily profitable.

1.2 Short-Vol ETP Complex: The Reflexivity Trap#

Two key instruments sat at the center of the blow-up:

• XIV (VelocityShares Daily Inverse VIX Short-Term ETN): Provided −1x daily exposure to the S&P 500 VIX Short-Term Futures Index.
• SVXY (ProShares Short VIX Short-Term Futures ETF): Similar −1x inverse daily exposure.

By January 2018, the AUM in inverse/short-volatility ETPs had grown to approximately $3–4 billion, with XIV alone holding ~$2 billion. Critically, these instruments rebalanced daily at the close by buying VIX futures when vol rose and selling when vol fell, creating a mechanical feedback loop:

$$ \Delta_{\text{hedge}} = \text{AUM} \times \frac{\Delta VIX_{futures}}{{VIX_{futures}}^2} $$

The notional vega exposure that needed to be rebalanced scaled quadratically with the size of the VIX move. A move from 14 to 28 (2x) required roughly 4x the rebalancing flow versus a move from 14 to 17.

2. The Intraday Mechanics of February 5, 2018#

2.1 Timeline and VIX Path#

2.2 The Gamma Squeeze: Quantitative Anatomy#

The VIX spike was not merely a volatility repricing — it was a convexity event driven by forced buying of VIX futures at an accelerating rate.

Step 1 — Initial Move. The S&P 500 dropped ~2.5% through the regular session, pushing VIX from ~17 to ~22. This alone was a standard 2-sigma equity move. VIX spot was tracking approximately:

$$ \Delta VIX \approx -\beta \times \Delta SPX \times \frac{1}{\sqrt{VIX_0}} $$

where β represents the vol-of-vol leverage and the inverse square-root term captures the convexity of the relationship at low VIX starting points.

Step 2 — Forced ETP Rebalancing. The inverse VIX ETPs needed to buy VIX front-month futures to rebalance their daily exposure. The approximate number of VIX futures contracts required was:

$$ N_{\text{contracts}} = \frac{\text{AUM} \times \Delta%{\text{VIX futures}}}{\text{Multiplier} \times P{\text{futures}}} $$

With AUM ~$3B, VIX futures up ~100%, and a contract multiplier of $1,000, this implied a rebalancing demand on the order of 100,000–200,000 VIX futures contracts — against a typical daily volume of ~300,000 contracts. This was an unabsorbable flow.

Step 3 — Feedback Loop. As the forced buying pushed VIX futures higher, the ETPs needed to buy even more contracts, because:

• Their short exposure grew larger with each tick higher
• The daily rebalancing formula demanded catching up to the closing level
• Other market participants (dealers, systematic funds) recognized the flow and front-ran it

This produced a positive convexity feedback in VIX — the more it moved, the more it had to move.

2.3 Term Structure Inversion#

The VIX futures curve, normally in contango (upward-sloping), violently inverted:

The M1–M2 spread swung from −1.16 (contango) to +3.26 (backwardation), a move of ~4.4 VIX points in a single session. The roll yield that had been funding short-vol strategies flipped to a massive headwind.

3. Impact on SPX Option Spread Pricing#

The VIX spike fundamentally altered the pricing of vertical spreads, calendars, and ratio structures across the S&P 500 options chain.

3.1 Implied Volatility Surface Deformation#

Parallel Shift (Vega)#

The at-the-money (ATM) implied volatility for the SPX 30-day tenor approximately doubled:

The front-end of the surface experienced the largest absolute shift, consistent with the standard term-structure of vega:

$$ \frac{\partial \sigma_T}{\partial \sigma_{\text{spot}}} \propto \frac{1}{\sqrt{T}} $$

Skew Steepening#

The 25-delta put skew (measured as 25Δ put IV minus ATM IV) steepened dramatically:

This skew steepening had asymmetric effects on put spreads vs. call spreads.

3.2 Vertical Spread Repricing#

Bull Call Spreads#

Consider a 30-DTE SPX bull call spread: long the 2750 call, short the 2800 call (with SPX at ~2700 post-selloff, both were slightly OTM).

Despite the SPX falling ~110 points (detrimental to a bullish position), the spread widened in absolute terms because the vega effect overwhelmed the delta effect. The long leg (closer to ATM) had higher vega than the short leg, and the massive vol expansion caused the long leg to gain more than the short leg.

The net greeks of the spread shifted:

$$ \Delta_{\text{spread}} \approx 0.20 \rightarrow 0.12 \quad (\text{delta compressed toward 0.5 for both legs}) $$

$$ \mathcal{V}{\text{spread}} = \mathcal{V}{2750C} - \mathcal{V}_{2800C} \approx +3.2 \text{ (net long vega)} $$

This net long vega position meant the spread appreciated even as the directional thesis (SPX higher) was failing.

Bear Put Spreads#

Consider a 30-DTE SPX bear put spread: long the 2700 put, short the 2650 put (now at-the-money and in-the-money respectively).

Bear put spreads experienced a double tailwind: (1) delta gains from SPX declining into the spread zone, and (2) net vega gains because skew steepening caused the long (higher-strike) put to gain vega faster than the short (lower-strike, deeper OTM) put would have in a parallel shift, but the skew effect partially offset this since the lower-strike put gained more IV. However, the delta/gamma contribution dominated:

$$ P&L_{\text{spread}} \approx \Delta_{\text{net}} \cdot \Delta S + \frac{1}{2}\Gamma_{\text{net}} \cdot (\Delta S)^2 + \mathcal{V}_{\text{net}} \cdot \Delta\sigma $$

For the bear put spread, all three terms were positive (net short delta × negative ΔS = positive; net long gamma × large move = positive; net vega impact depended on skew dynamics but was generally positive).

3.3 Calendar Spread Destruction#

Calendar spreads (long back-month, short front-month at same strike) experienced severe mark-to-market losses. The key driver was the collapse in the implied volatility term structure ratio.

Example: ATM calendar spread, long April / short February at 2750 strike:

The calendar lost ~64% of its value because the front-month vol exploded 3x while the back-month only rose ~1.7x. Since the trader was short the front-month (which had the highest vega sensitivity per unit of vol), the short leg’s mark-to-market loss overwhelmed the long leg’s gain.

The mathematical driver:

$$ \mathcal{V}{\text{calendar}} = \mathcal{V}{\text{back}} - \mathcal{V}_{\text{front}} $$

Since $\mathcal{V} \propto \sqrt{T}$ and the front-month vol rose disproportionately, the net vega of the calendar (normally positive, benefiting from vol expansion) flipped effective sign due to the non-parallel shift.

3.4 Iron Condor and Credit Spread Blow-Up#

Short premium strategies were devastated. A typical 30-DTE SPX iron condor positioned at ±1.5 standard deviations:

The put side alone generated a loss of ~7.3x the credit received. The critical issue was that the put spread, initially ~1.5σ OTM, was now ATM/ITM, and the vol expansion massively inflated the residual extrinsic value of both legs.

The maximum theoretical loss on the 20-point-wide put spread was $20.00. The actual mark-to-market loss approached this maximum, despite 25 DTE remaining, because:

1. Delta of the short 2680 put went from ~0.18 to ~0.55
2. Gamma concentration near ATM accelerated the delta shift
3. Vega expansion added ~$6–8 of extrinsic value back into both legs, but the short leg (now closer to ATM) gained more

4. Second-Order Effects and Dealer Hedging#

4.1 Dealer Gamma Exposure#

SPX options market makers were estimated to be long gamma in aggregate entering the event (due to systematic overwriting strategies by asset managers creating net short call positions for dealers to absorb). As SPX fell, dealers needed to sell futures to delta-hedge, amplifying the selloff.

The total dealer gamma notional was estimated at:

$$ \Gamma_{\text{notional}} = \sum_i \Gamma_i \times OI_i \times 100 \times S^2 $$

At an aggregate $\Gamma_{\text{notional}}$ estimated near $5–10 billion per 1% SPX move, the 4.1% decline forced an estimated $20–40 billion in futures selling from dealer hedging alone.

4.2 Vanna and Volga Effects#

The simultaneous rise in volatility and decline in spot triggered vanna flows (cross-partial of delta with respect to vol):

$$ \text{Vanna} = \frac{\partial \Delta}{\partial \sigma} = \frac{\partial \mathcal{V}}{\partial S} $$

For put options, vanna is typically positive (as vol rises, put delta becomes more negative), requiring dealers who were short puts to sell additional futures — another amplifying flow.

Volga (sensitivity of vega to vol) caused the cost of hedging vol-of-vol exposure to spike, widening bid-ask spreads on options to 2–5x normal levels, further reducing liquidity.

5. Product Termination and Aftermath#

5.1 XIV Termination Event#

Credit Suisse’s XIV ETN contained an acceleration clause: if the indicative value declined more than 80% in a single day, the issuer could elect to terminate the product. On February 5, the indicative value fell approximately 96% from $99 to ~$4. Credit Suisse announced termination on February 6, with final redemption at ~$5.99 per share on February 20.

5.2 SVXY Restructuring#

ProShares reduced SVXY’s exposure from −1x to −0.5x daily VIX short-term futures, effectively halving the leverage and the rebalancing feedback.

5.3 Lasting Market Structure Changes#

The event demonstrated several quantitative truths about volatility markets:

1. Short vol carry strategies have fat-tailed loss distributions: The Sharpe ratio of ~3.0 observed pre-event was an artifact of understated tail risk, not genuine alpha.
2. Daily rebalancing of leveraged/inverse ETPs creates endogenous fragility: The rebalancing flow scales non-linearly with the size of the move, creating a reflexive amplification mechanism.
3. VIX term structure is not arbitrage-free under stress: The carry earned from contango assumes an orderly reversion path; in practice, the roll yield can be eliminated in a single session.
4. Option spread pricing models assuming parallel vol shifts fail catastrophically: The non-uniform surface deformation (front-loaded, skew-steepening) creates P&L outcomes for spreads that are unintuitive under standard Black-Scholes partial-derivative Greeks.

6. Key Quantitative Takeaways#

This analysis reconstructs the event using standard options pricing mathematics, published VIX settlement data, and structural features of the ETP rebalancing mechanism. Specific spread examples use representative pricing consistent with the observed surface deformation but should not be treated as exact transaction-level data.