Every day, dealers who sell zero-days-to-expiry SPX options must hedge their positions in real time. When their aggregate gamma exposure turns negative, that hedging mechanically amplifies intraday moves. We measure this effect empirically using OptionMetrics and TAQ data across 887 trading days.
Zero-days-to-expiry (0DTE) SPX options, contracts that expire the same calendar day they are traded, have grown from a curiosity to a dominant force in US equity markets. Since the CBOE introduced daily SPX expirations in May 2022, 0DTE options now routinely account for more than 40% of total SPX options volume. Behind every one of those contracts is a market maker who has sold it and must now manage the resulting risk.
When a dealer sells an option, they delta-hedge their exposure by taking an offsetting position in the underlying. As the underlying price moves, delta changes, and the dealer must adjust their hedge continuously. The rate at which delta changes with price is gamma (Γ). A dealer who is long gamma profits from this adjustment process: they buy when prices fall and sell when prices rise, acting as a natural stabiliser. A dealer who is short gamma must do the opposite: buy as prices rise and sell as prices fall. Their hedging amplifies the very moves they are trying to hedge.
Dealer Gamma Exposure (GEX) aggregates this effect across all open 0DTE contracts. Under the standard assumption (that all open interest represents customer-bought positions with dealers on the other side), GEX is computed as:
The call-minus-put sign convention reflects empirical observation: SPX call open interest is predominantly driven by customer hedgers and income sellers, while put open interest is dominated by protective buyers. Dealers tend to be net short calls and net long puts in aggregate, yielding positive net gamma under normal conditions. But 0DTE options are different: they expire the same day, so open interest at the start of the session represents only same-day speculative and hedging activity, with a sharper intraday gamma profile than any multi-day contract.
We pull every SPX 0DTE option record from OptionMetrics via WRDS for the period Jan 2022 to Aug 2025. A contract qualifies as 0DTE if its expiration date equals the quote date. For each trading day, we sum the signed gamma × OI product across all qualifying strikes and apply the GEX formula above, using the SPY closing price as a proxy for the SPX spot price in dollar-normalisation.
Realized intraday volatility is computed from TAQ consolidated trades for SPY. For each day, we aggregate individual trades into five-minute VWAP bars from 9:30 am to 4:00 pm EST (78 intervals), compute the sequence of log returns, and annualise the realised variance: σ = √(∑ r ² × 252 × 78). This measure captures intraday price variability without contamination from overnight gap risk.
Over the full sample, 53% of trading days had positive GEX, meaning dealers were net long gamma on the majority of sessions. The distribution is right-skewed: large positive GEX readings (dealers very long gamma) are more common than deeply negative ones. Negative GEX days tend to cluster around macro events, FOMC announcements, and periods of elevated VIX, when put buyers dominate 0DTE flow.
We classify each trading day into one of three regimes based on daily GEX: Negative GEX (gex < 0), Low GEX (0 ≤ gex < median of positive days), and High GEX (gex ≥ median of positive days). The key comparison is between the extremes.
The vol premium is statistically significant. On negative-GEX days, mean annualised intraday realised volatility is 13.7%, compared with 8.1% on high-GEX days, a relative vol premium of roughly 70%. This is not merely a reflection of regime selection bias (e.g., negative GEX days happening to coincide with scheduled macro events). The chart below shows the distribution of intraday vol across all three regimes.
The scatter plot reveals a clear negative relationship: higher GEX is associated with lower intraday vol. The OLS regression has R² = 0.049, meaning GEX explains approximately 5% of the cross-day variation in intraday realised vol. That is a small number. The remaining 95% of day-to-day vol variation is driven by other factors: scheduled macro events, the prevailing VIX regime, earnings clusters, and seasonal patterns. GEX alone cannot forecast intraday vol with precision. What it does offer is a reliable regime-level signal: the difference in group means between negative and high GEX days is large (+70%) and statistically robust (t = 11.95, p < 0.0001). That is a coarse signal, useful for adjusting position size at the start of the session, not a precise vol forecast.
GEX is not uniform through the trading day. The 30-minute bucket data shows two pronounced peaks in the regime divergence: the 10:00 bucket (Negative GEX: 141, High GEX: 52, spread: 89 percentage points) and the 13:00 bucket (Negative GEX: 148, High GEX: 60, spread: 89 percentage points). The “charm” effect (dΔ/dt) and intraday gamma acceleration both contribute, but the largest vol separation occurs well before the final hour rather than concentrating at the close.
To capture the intraday pattern, we split each trading day into 30-minute buckets (9:30, 10:00, …, 15:30) and compute the average realised vol within each bucket, grouped by GEX regime.
The divergence between regimes peaks at the 10:00 and 13:00 buckets, where Negative GEX realised vol reaches 141 and 148 respectively against High GEX levels below 60. The spread narrows through mid-afternoon (the 15:00 bucket shows the smallest gap of the session) before partially recovering at 15:30, the final data point. On Negative GEX days, dealers face compounding hedge requirements as positions accumulate through the session, amplifying vol across the full trading day rather than concentrating it exclusively near the close.
The morning (9:30 – 10:00) also shows an elevated vol spread between regimes. This reflects the opening print dynamics: on negative-GEX days, dealers entering the session with inherited short-gamma positions from overnight activity must hedge aggressively into the open, when bid-ask spreads are widest.
The GEX regime classification translates directly into a practical position-sizing rule. The logic is simple: if you know that negative-GEX days have systematically higher intraday vol, you can scale intraday position sizes inversely to the vol forecast. Rather than adjusting to a point forecast, we use the regime median as a baseline.
The time-series chart confirms the pattern holds out-of-sample throughout the period; negative-GEX days (shown in red) consistently cluster around the higher end of the daily vol distribution. The effect is most visible during 2022 – 2023, when the combination of Fed rate hikes and elevated VIX produced frequent negative-GEX days. During the calmer stretches of 2024–2025, positive GEX dominated and intraday vol remained suppressed for extended stretches.
Several important limitations apply to the findings above.
The standard GEX assumption is a simplification. We assume all open interest is customer-bought. In practice, dealers take two-sided positions, and some customers are net sellers of 0DTE options (income-generating strategies, institutional hedges). Without signed order-flow data at the individual trade level, we cannot construct a precise estimate of net dealer gamma. The bias introduced by this assumption is likely to overstate the magnitude of GEX swings but should preserve the directional signal.
Correlation is not causation. The relationship between negative GEX and elevated vol could partly reflect reverse causation: high-vol days attract more put buyers (generating negative GEX), rather than negative GEX causing higher vol. Disentangling this would require intraday GEX snapshots taken at the open, before the day’s vol is realised, a data exercise beyond the scope of this study.
GEX flips intraday. We use beginning-of-day GEX computed from OptionMetrics daily snapshots. In reality, large trades or sharp price moves can shift the aggregate GEX level substantially within a session. A more sophisticated implementation would use intraday GEX updates from a real-time options feed.
Zero-days-to-expiry options have fundamentally changed the microstructure of the SPX. The daily flow of 0DTE gamma from dealers to customers, and the resulting obligation on dealers to delta-hedge in real time, creates a structural, measurable regime effect on intraday volatility. When dealer gamma is positive, they act as a dampener on price moves. When it turns negative, they become amplifiers.
Across 887 trading days from Jan 2022 to Aug 2025, negative-GEX days experienced mean intraday realised vol of 13.7%, compared with 8.1% on high-GEX days. The difference is statistically significant (p = 0.0000), and the pattern is persistent through time rather than driven by a handful of outlier sessions. The regime divergence peaks at the open (10:00 bucket) and in the early afternoon (13:00 bucket), where the Negative-to-High GEX vol spread reaches 89 percentage points in both windows. The mid-afternoon narrows before a partial recovery at the close, confirming that the effect is persistent across the full session rather than concentrated in the final hour.
For practitioners, this translates into a simple, computable signal: check the sign of aggregate 0DTE GEX before the open. On negative-GEX days, the market is more likely to exhibit volatile, whipsaw intraday price action. Widen stops, reduce intraday position sizes, and treat apparent support and resistance levels with greater scepticism.