The February 2026 Selloff › The Machinery That Amplifies Selling

The math explains why losses are harder to recover from. The market's infrastructure explains why they happen so fast. The February 2026 selloff was not driven by millions of individual investors independently deciding to sell. Interlocking mechanical systems drove it, each responding to price signals in ways that amplified the decline. Start with volatility targeting and risk parity. An estimated \$2 trillion globally sits in volatility-targeting strategies, with another \$300 billion in risk parity funds. These strategies target a set portfolio volatility. When realized volatility is low, they add leverage and equity exposure. When volatility spikes, they sell to reduce risk. During March 2020, a stylized risk parity portfolio had to sell assets worth nearly 225% of its capital to meet its volatility target. During the August 2024 selloff, estimates suggested systematic strategies would dump \$70 to \$80 billion in equities in a single day, with at least \$90 billion more to follow over subsequent sessions. These are automatic responses hardwired into code. Then there is the gamma trap. The options market has become a major amplifier of equity moves. Over 50% of S&P 500 options volume now comes from same-day expiry contracts (0DTE options that barely existed before 2022). These contracts carry extreme gamma: small movements in the underlying index create enormous shifts in the delta that dealers must hedge. When the market sits in "positive gamma" territory, dealers are naturally long options, and their hedging dampens volatility, buying dips and selling rallies. When the market enters "negative gamma" territory, the dynamic reverses. Dealers who are short options must sell as markets fall and buy as they rise, amplifying moves in both directions. SPX options carry approximately \$80 billion in gross gamma exposure. A 1% move in the index translates to a delta shift of roughly \$80 billion, all of which must be hedged. When that hedging is procyclical, prices fall, dealers sell to hedge, prices fall further, dealers sell more. The interaction between gamma, vanna, and charm makes this worse during selloffs. Vanna measures how an option's delta changes with implied volatility. When volatility surges, the deltas on short put positions climb, forcing dealers to sell more of the underlying asset. Charm, which measures how delta changes with time, normally provides a gentle buying pressure as out-of-the-money puts decay. During sharp selloffs, charm flows are overwhelmed by gamma and vanna effects, and the options market stops absorbing shocks and starts causing them. Margin calls add fuel. As of December 2025, FINRA-reported margin debt stood at \$1.23 trillion, a seventh consecutive record high. Relative to GDP, margin debt hit 3.91%, well above both the dot-com bubble peak of 2.6% and the 2007 pre-crisis peak of 2.5%. When prices fall, portfolios dip below maintenance margin requirements, brokers issue margin calls, and investors must deposit cash or liquidate. That forced selling drives prices lower, which triggers more margin calls, which forces more selling. Research using Indian stock market data found margin trading creates "excess comovement" during crises, meaning stocks fall together more than fundamentals warrant, driven largely by brokers liquidating concentrated client positions. This comovement effect appears only during selloffs, never during rallies. There is no equivalent forced-buying mechanism that creates upward cascades. Finally, algorithmic trend-following. CTA and managed futures strategies, managing roughly \$318 billion globally, follow price trends across asset classes. When prices breach technical levels, these algorithms sell. During February 2026, different systematic strategies deleveraged at different speeds: momentum strategies reacted within hours, while volatility-targeting funds adjusted over days as trailing realized volatility calculations incorporated the new data. The staggered nature of these unwinds means systematic selling persists for days or weeks after the initial shock, producing the "slow bleed" pattern that characterized February. The combined effect is sometimes called a "doom loop": dealer hedging flows trigger CTA momentum signals, which propagate negative price action, which causes volatility-targeting funds to deleverage, which increases realized volatility, which forces risk parity unwinds, which pushes cross-asset correlations toward one. Each mechanism feeds into the next. Each was designed independently and rationally. Together, they produce synchronized selling across asset classes.