Event Volatility: Total Variance, Jump Residuals, and Executable Risk

For educational purposes only; not investment advice. Investing may result in loss.

Direct answer

Event volatility is a model-dependent price of uncertainty associated with a scheduled catalyst, such as earnings, a policy decision, court ruling, clinical result, election, or economic release. Options that expire after the event can carry more total implied variance than comparable options expiring before it, but the difference also contains ordinary variance, overlapping events, skew, risk premium, quote noise, and model error.

A scalar event sigma has no necessary direction and is not an arithmetic expected move, probability, quantile, or physical forecast. The full strike-expiration surface can imply asymmetric or multimodal jump risk. Actual trade P/L additionally depends on the structure, entry and exit sides, spot jump, post-event surface, time, carry, fees, liquidity, and lifecycle.

Build the variance decomposition

At forward log-moneyness coordinate k and year fraction T, a Black-Scholes-equivalent total implied variance coordinate is:

W(k,T) = sigma_imp(k,T)^2 x T

For synchronized comparable coordinates at T_1 < T_2:

v_forward = [W(k,T_2) - W(k,T_1)] / (T_2 - T_1)

This is the annualized variance for the entire interval under the chosen quote and model conventions. It is not automatically the variance of one scheduled event. Under an explicit additive log-return variance and zero-covariance approximation, a conditional residual can be written:

v_event = W(k,T_2) - W(k,T_1) - v_base x (T_2 - T_1) - sum(v_other,j)

Only a validated nonnegative residual supports:

event sigma = sqrt(v_event)

A negative or unstable residual is a data, coordinate, baseline, surface, or model diagnostic; do not silently floor it to zero. Event sigma is a risk-neutral model-implied log-return standard deviation under the stated assumptions. The arithmetic markers exp(event sigma) - 1 and 1 - exp(-event sigma) are asymmetric and still are not confidence bounds.

Quotes must share timestamp, forward, moneyness or Delta convention, rates, dividends or borrow, exercise treatment, day count, and interpolation. Same literal strike, spot-ATM strike at each expiration, and Delta-ATM are not interchangeable. American single-stock implied volatility is itself model-dependent because early exercise and distributions matter.

Four worked examples

• Interval forward variance. Let T_1 = 7 / 365, sigma_1 = 30%, T_2 = 14 / 365, and sigma_2 = 50%. Then W_1 = 0.001726027397, W_2 = 0.009589041096, and the increment is 0.007863013699. Interval forward variance is 0.410000000000, so annualized interval forward volatility is 64.0312423743%. It covers every risk in the intervening seven days.
• Baseline-dependent event residual. Let T_1 = 10 / 365, sigma_1 = 30%, T_2 = 17 / 365, sigma_2 = 46%, and ordinary interval baseline sigma_base = 28%. Total variances are 0.002465753425 and 0.009855342466; the increment is 0.007389589041, and ordinary seven-day variance is 0.001503561644. Thus v_event = 0.005886027397 and event sigma = 7.6720449668%. Symmetric one-sigma log markers map to arithmetic +7.9740192469% and -7.3851277396%, not probability bounds or expected moves.
• Surface coordinate matters. For seven- and fourteen-day expirations, ATM IVs of 25% / 40% give W_1 = 0.001198630137, W_2 = 0.006136986301, interval forward variance 0.2575, and forward volatility 50.7444578255%. At matched 90% forward moneyness, put IVs of 35% / 50% give W_1 = 0.002349315068, W_2 = 0.009589041096, forward variance 0.3775, and forward volatility 61.4410286372%. One ATM scalar does not describe the tail.
• Executable price ledger. At S_0 = $100, one M = 100 straddle has call $3.60 bid / $4.10 ask and put $3.40 bid / $3.90 ask. The leg-summed buy reference is m_buy = 8.00%, the sell reference is m_sell = 7.00%, and the immediate displayed width is $100 per package before fees, without a package-fill guarantee. Long opening cash is -$800. After the event, S_1 = $106 and an executable package bid of $6.35 creates +$635 exit cash and -$165 P/L before fees despite the +6% spot move.

Seven-step workflow and controls

1. Freeze the authoritative event source, expected and actual release instant, time zone, session, revision, delay or leak status, and overlapping catalysts.
2. Define the legal option class, exercise style, settlement, multiplier, last trade, expiration and two maturities that truly exclude and include the event.
3. Capture synchronized bid, ask, size and package quotes, spot, forward inputs, rates, dividends or borrow, and consistent strike, Delta or forward-moneyness coordinates.
4. Clean and fit an arbitrage-aware implied-volatility surface, convert comparable coordinates to total variance, and retain bid-ask and interpolation sensitivity rather than one midpoint.
5. Specify and version the ordinary baseline and other events; calculate interval forward variance and the conditional event residual; flag negative or unstable outputs rather than clipping them.
6. Translate the model into executable long, short, calendar or hedge structures, then stress full-surface jump paths, gaps, halts, partial fills, margin, assignment and frozen or discrete hedges.
7. After the event, reconcile option, stock and cash fills; attribute full-repriced P/L to spot, surface, time, carry and costs; retain residual; and compare implied with realized log-return evidence without look-ahead.

• The event source, date, actual time, time zone, or release session can be wrong.
• An announcement can be delayed, revised, leaked, staged, or followed by another release.
• Last-trade, expiration or official-settlement clocks may fail to bracket the event.
• Calendar-day, trading-day, minute and annualization conventions can be mixed.
• Spot, forward inputs and option quotes can be stale or asynchronous.
• Subtracting raw annualized implied volatilities confuses volatility with time-scaled variance.
• Literal strike, spot ATM, forward ATM and Delta coordinates can identify different surface points.
• American exercise, dividends, borrow or cash-versus-physical settlement can distort model comparison.
• The ordinary baseline can be unstable, regime-dependent or selected with hindsight.
• Overlapping catalysts and covariance can be omitted from an additive decomposition.
• A noisy or arbitrage-inconsistent surface can create negative forward or event residual variance.
• One ATM scalar can hide skew, asymmetric tails, multimodality and strike-specific liquidity.
• Wide, stale or zero-bid markets and insufficient size can make implied volatility unstable.
• Leg-summed bids and asks do not guarantee a simultaneous package fill.
• Straddle-to-spot ratios and event sigma can be mislabeled as expected moves or probabilities.
• Jump direction, tail thickness and the risk-neutral versus physical distribution can be confused.
• Post-event IV decline does not guarantee a short-volatility profit.
• A Delta hedge or stop cannot trade through an overnight gap, halt or closed market.
• Discrete hedging, cross Greeks, surface dynamics and higher orders can create full-repricing residuals.
• Margin, assignment, settlement, fees, taxes, data vintage, small samples and regime shifts can invalidate the result.

Common misconceptions

• “Subtracting two annualized IVs isolates the event.” Comparable total variance and a disclosed baseline are required.
• “Event sigma predicts direction, probability, or the arithmetic expected move.” It is a model-implied log-return standard deviation under assumptions.
• “A straddle midpoint divided by spot is the event’s true expected move.” Quote side, maturity, skew, non-event time and execution matter.
• “IV crush guarantees option sellers profit and stops cap the gap.” The jump and unavailable execution can dominate.
• “One event has one clean volatility independent of strike, expiry, baseline and other catalysts.” The surface and decomposition are conditional.

Related topics

• Implied volatility
• Earnings straddle expected move
• Volatility surface

Primary sources

• Cboe Volatility Index Mathematics Methodology
• VIX Term Structure
• U.S. Election 2024—What the Options Data Says
• Characteristics and Risks of Standardized Options
• Option Price Behavior
• Option Pricing When Underlying Stock Returns Are Discontinuous
• Anticipated Information Releases Reflected in Call Option Prices
• Option Pricing of Earnings Announcement Risks