Implied Volatility in Options: Pricing, IV Crush, and Scenarios

Editorial responsibility: Oliver Sinz · Last verified: October 26, 2026

Implied volatility (IV) is the volatility input that makes an option pricing model reproduce an option’s observed price. Nobody measures it directly. You take a premium from the market, fix the stock price, strike, time to expiration, interest rate, and dividends, and solve for the one volatility the model needs to produce that premium. The result is a price expressed in volatility units: not a measurement of how much the stock has moved, and not a guaranteed forecast of how much it will move.

That distinction explains most of what confuses traders about IV: why two calculators show different numbers for the same quote, why index options have tended to imply more volatility than was later realized, why options can lose value after earnings even when the stock moves the right way, and why a scenario’s answer depends on what you assume IV does next.

All prices, implied volatilities, and scenarios in this article are hypothetical model calculations, not market data. Nothing here is a recommendation to trade.

What implied volatility measures

A pricing model turns six inputs into a value: underlying price, strike, time to expiration, interest rate, dividends, and volatility. The strike and expiration are contractual and the underlying price is observable. The rate and future dividends usually involve some assumption, estimate, or inference from other market prices. Volatility, the annualized standard deviation of the underlying’s log returns over the option’s life, is the one input that cannot be observed in advance. Implied volatility runs the model in reverse. Researchers began backing volatilities out of option prices soon after the Black and Scholes (1973) and Merton (1973) models appeared; Latané and Rendleman (1976) is an early example.

Three consequences follow:

• IV belongs to a model and a set of inputs. Under Black–Scholes–Merton (European exercise, constant volatility, lognormal prices), each option price maps to one IV. Change the model, the dividend assumption, or the day count, and the same premium implies a different number.
• IV is specific to each option. Every strike and expiration has its own. “The stock’s IV” on a quote screen is a summary, usually of near-the-money options at one expiration, or an index built from many options.
• IV is a price, so it carries everything that moves prices. That includes expectations of future movement, but also compensation for bearing volatility risk, demand for hedges, and scheduled events.

What an implied volatility number means

IV is quoted per year. To translate it to the option’s horizon, multiply by the square root of time in years. For a hypothetical stock XYZ at $100 and a 45-day option at 25% IV:

• Move to expiration: 0.25 × √(45/365) ≈ 8.78%, one standard deviation of the log return over 45 days.
• Daily move: 0.25 ÷ √365 ≈ 1.31% per calendar day, or 0.25 ÷ √252 ≈ 1.57% per trading day (the “divide by 16” shortcut, since √256 = 16).

Because the model treats returns as lognormal, the one-standard-deviation price range is lopsided. Centered on the model’s risk-neutral median, $100 × e^((r − σ²/2)T) with a 4% risk-free rate, it runs from about $91.70 to $109.29 for the 45-day option. For a one-year option at 30% IV, it runs from $73.71 to $134.31, down 26% and up 34%, not “plus or minus 30%.” At 20%, the one-year range is $83.53 to $124.61. Under the model’s own assumptions, the stock ends inside its range about 68% of the time. That is a risk-neutral probability built into the price, not a real-world forecast, and option prices across strikes typically depart from this single-volatility benchmark, as the skew section shows.

There is no universally “high” or “good” IV. The same 30% can be unusually high for one underlying and unusually low for another, so platforms compare IV with its own history. IV rank is commonly defined as (current IV − 52-week low) ÷ (52-week high − 52-week low), and IV percentile as the share of days in the past year with lower IV. Definitions and data vary by platform, and neither metric says whether options are cheap or expensive relative to the volatility that will actually occur.

How implied volatility is calculated

There is no simple closed-form implied volatility formula for the Black–Scholes–Merton model; in practice, IV is solved numerically. The search works because, for European options, the model value rises strictly with volatility, so at most one volatility matches a given price. A solution exists only when the price lies within no-arbitrage bounds. For a call on a stock without dividends, that means above max(0, S − PV(K)), where PV(K) is the strike discounted at the risk-free rate, and below the stock price S. Manaster and Koehler (1982) set out the conditions under which a positive implied variance exists. For the 45-day 100 call, any quote at or below $0.49 or at or above $100 has no implied volatility under this model.

A common approach is Newton–Raphson iteration, which uses vega (the price change per unit of volatility) as the slope, with bisection or another bracketing method as a fallback where vega is tiny: deep in or out of the money, or close to expiration. A convenient starting guess comes from Brenner and Subrahmanyam (1988): an option struck at the forward price is worth about 0.4 × S × σ × √T, so σ ≈ √(2π) × C ÷ (S × √T), where C is the call price and T the time in years.

A worked inversion

Hypothetical quote: a 45-day XYZ 100 call bid at $4.40 and offered at $4.60, midpoint $4.50. Other inputs: XYZ at $100.00, 4.0% continuously compounded risk-free rate, no dividends, T = 45/365 = 0.1233.

1. Starting guess: √(2π) × 4.50 ÷ (100 × √0.1233) ≈ 32.12%.
2. First iteration: at 32.12%, the model value is $4.7361, $0.2361 too high. Vega is about $0.1394 per volatility point, so the next guess is 32.12% − (0.2361 ÷ 0.1394) points ≈ 30.426%.
3. Second iteration: at 30.426%, the model value is $4.5000, within a thousandth of a cent of the target. The implied volatility is 30.43%.

The shortcut overshot because the $100 strike sits below the $100.49 forward price, so part of the premium reflects the forward rather than volatility. Newton’s method converged almost at once because an at-the-money option’s value is nearly linear in volatility.

The bid and the ask imply 29.71% and 31.14%, so a $0.20-wide market spans 1.4 volatility points. For cheap out-of-the-money options the problem is much worse. A 45-day 120 call quoted $0.05 bid, $0.10 offered implies 23.4% to 26.1%: a five-cent spread spans 2.6 points. Per one cent of spread, that IV range is about seven times as wide as the at-the-money call’s.

Why implied volatility calculators disagree

The same option can produce noticeably different IVs depending on the inputs and model. The first three rows use the $4.50 call above. The last four use hypothetical prices generated with a “true” volatility of 25%; the dividend rows subtract the dividend’s present value from the stock price.

Three lessons sit in this table:

• A gap between call and put IVs at the same strike often signals an input problem, such as a missing dividend, a borrowing cost, or the wrong rate, or the early-exercise value of American options, rather than two different views of volatility. Deriving the forward price from the options themselves sidesteps a separate dividend forecast; Cboe’s volatility indexes, for example, take it from put-call parity at the strike where call and put prices are closest (Cboe Volatility Index Mathematics Methodology).
• Exercise style matters. Standard U.S. equity options are American-style (OIC: Options Exercise FAQ). A European formula attributes the early-exercise premium to volatility, so it overstates the IV of in-the-money puts. The American model values here come from a binomial tree (Cox, Ross, and Rubinstein, 1979).
• An American option trading at its exercise value has no well-defined IV. The deep in-the-money American put is worth exactly $20.00 at any volatility up to about 28%, so its price cannot identify one.

Implied vs. realized volatility

Realized, or historical, volatility is computed from past prices. With daily log returns rᵢ = ln(Pᵢ ÷ Pᵢ₋₁) over n days, a standard estimate is √[252 ÷ (n − 1) × Σ(rᵢ − r̄)²]. Conventions differ in the window (10, 20, or 60 days), close-to-close versus intraday data, whether the mean is subtracted, and the annualization factor, so two “historical volatility” figures for the same stock can disagree.

The comparison also needs matching horizons. A 45-day option’s IV refers to the next 45 days, so the relevant benchmark is volatility realized over those same days, which is unknown when the option trades. Trailing 20-day realized volatility covers a different window; ahead of earnings, it may contain no announcement while the option’s life contains one.

The gap between the two is what a delta-hedged option position trades. Hedged against small moves, a long option tends to gain when the stock moves more than its IV implied and lose when it moves less, before costs; a short option has the reverse exposure. In index options, buyers have historically paid a premium for that exposure.

Carr and Wu (2009) show that the risk-neutral expected variance, also called the variance swap rate, can be synthesized from a portfolio of options. They measure the variance risk premium as realized variance minus that rate across five stock indexes and 35 individual stocks and find it strongly negative for the S&P 500, S&P 100, and Dow Jones Industrial Average. For the Nasdaq-100 and most individual stocks it was also negative but smaller in magnitude, and often statistically insignificant for individual stocks. Bakshi and Kapadia (2003) found that delta-hedged long S&P 500 index option positions lost money on average. Both results describe specific markets and samples, not a rule for any stock or period, and short-volatility positions carry the risk of large losses when volatility jumps.

Is implied volatility a forecast?

Only partly. For S&P 100 index options, Canina and Figlewski (1993) found IV to be a poor forecast of subsequent realized volatility, with virtually no correlation and no incorporation of recently observed volatility. Using longer time series and non-overlapping data, Christensen and Prabhala (1998) found that IV outperformed past volatility in forecasting future volatility, and attributed earlier evidence of bias to a regime shift around the October 1987 crash. A fair reading: IV contains information about future volatility, but it is biased by risk premiums, noisy across strikes, and dependent on the model and options used to extract it.

VIX: implied volatility across many options

The Cboe Volatility Index (VIX) is designed to measure the 30-day volatility expectation conveyed by S&P 500 index option prices. It does not invert Black–Scholes for any single option. Its methodology rests on variance swap pricing: it combines out-of-the-money puts and calls across strikes, weights each by its strike interval divided by the strike squared, measures time in minutes of a 365-day year, and interpolates between two expirations to a constant 30-day term (Cboe Mathematics Methodology). A VIX of 20 reads like 20% annualized volatility, but it is not the IV of any one option, nor the “IV” a broker shows for an individual stock.

How implied volatility changes reprice options

Holding the stock at $100 and the rate at 4%, the table reprices five hypothetical options at different IVs:

These are European-style model values. An American-style 90 put would be worth slightly more, about $0.42 at 25%.

• At the money, value moves roughly in proportion to IV. A 5-point change from 25% moves the premium by close to a fifth (16% to 19% here) at every expiration, because an at-the-money option is worth approximately 0.4 × S × σ × √T. In dollars, the effect grows with time: about $1.91 for the one-year call against $0.28 for the 7-day call.
• Out of the money, the response is convex and lopsided. The 90 put gains 83% on a 5-point rise and loses 60% on a 5-point drop.
• Multi-leg positions add up very uneven leg changes. A position’s net vega is the sum of its legs’ vegas and describes a parallel shift, with every leg’s IV moving by the same number of points. Real IV changes are rarely parallel across strikes or expirations, and scheduled events produce the clearest example.

IV crush: why option prices fall after scheduled events

An option expiring after a scheduled announcement, such as an earnings report, must price the chance of a large one-day move. The simplest models of this kind treat the announcement-day return as a normally distributed jump on a known date. A European option is then priced on the total variance to expiration, so its implied variance is an average over the option’s life:

σ_IV² × T = σ_base² × T + σ_event²

Here σ_base is the volatility on ordinary days and σ_event is the standard deviation of the log return on the announcement day. The event term is a fixed amount of variance. Spread over a short remaining life, it raises IV sharply; spread over a long one, only a little. Once the news is out, the term drops out of every expiration at once. That is IV crush: a known block of variance leaving the option’s remaining life, not merely a change in sentiment.

Research supports this structure. Patell and Wolfson (1979) documented systematic changes in Black–Scholes implied variances around annual earnings announcements. Dubinsky, Johannes, Kaeck, and Seeger (2019) developed estimators that separate earnings-announcement uncertainty from normal day-to-day volatility. They found the anticipated uncertainty quantitatively large, varying over time, and informative about future return volatility.

Separating the event from normal volatility

Hypothetical setup: XYZ at $100 reports after today’s close. The front expiration, 8 days away, trades at 50.4% IV; the next one, 36 days away, at 35.6%. Assume both contain the same single event and the same base volatility, and measure time in calendar days ÷ 365. The two IVs give two equations in two unknowns:

• Base volatility: σ_base² = (σ₂² × T₂ − σ₁² × T₁) ÷ (T₂ − T₁) = (0.01250 − 0.00557) ÷ (0.09863 − 0.02192) = 0.0904, so σ_base ≈ 30.1%.
• Event volatility: σ_event² = σ₁² × T₁ − σ_base² × T₁ = 0.00557 − 0.00198 = 0.00359, so σ_event ≈ 6.0%.

If base volatility is unchanged after the report, both expirations fall to about 30%: the front month loses about 20 IV points, the back month about 5.5. A trading-day convention, which assigns little or no variance to weekends, would change these numbers, so the decomposition is only as good as its time convention and its assumption that nothing else differs between the two expirations.

The implied move

The event’s 6.0% standard deviation corresponds to an expected absolute move of about √(2/π) × 6.0% ≈ 4.8%. The front-month at-the-money straddle costs $5.95 at 50.4% IV, and a “straddle price ÷ stock price” rule would call that a 5.95% implied move. The straddle also prices eight days of ordinary volatility, however: its value matches the expected absolute move of the event and those eight days combined, about 5.95%. The farther the expiration from the event, the more a straddle-based “implied move” overstates the move priced for the event itself. Both figures are risk-neutral quantities, not predictions.

A long call held through the announcement

Suppose a trader buys the 8-day 100 call at 50.4% IV for $3.02, or $302 per contract. The next day, with 7 days left, the table shows profit or loss per contract by stock price and by the IV after the report. Here and in later scenario tables, P&L uses unrounded model values while premiums are shown rounded, so recomputing from the displayed premiums can differ by up to about $1 per contract.

If IV falls to the 30% base level, the call needs XYZ above about $102.15 the next day to break even; with IV unchanged, about $100.37 would be enough. A rise of up to about 2% is “right on direction” and still a loss. Nothing guarantees a crush, though. The 30% column is what the pre-event term structure implies if base volatility stays where it was. If the report leaves more uncertainty than before, IV can stay elevated or rise.

How much did the crush cost?

At $104 with IV at 30%, the call gained $144 per contract. Splitting that into price, time, and volatility effects requires repricing one input at a time, and the answer depends on the order:

Each row adds up to +$144. The volatility effect is smaller when the price move is applied first, because the call is then $4 in the money and less sensitive to volatility. The difference is the interaction between inputs that moved together; no single attribution is correct, so a report should state its convention or use a symmetric average such as the last row.

Spreads and calendars: the crush is not parallel

The same event, repriced after the report at 30% IV for every option, affects positions very differently. Profit or loss per contract:

The 36-day call is priced at 35.6% before the report, the spread and calendar costs are net debits, and the vegas below are measured at purchase, with XYZ at $100 and the pre-event IVs.

• More vega does not mean more crush. The 36-day call has about twice the vega of the 8-day call ($12.48 against $5.90 per point per contract) yet loses less at an unchanged price ($76 against $132). Its IV falls about 5.5 points instead of 20, and one day of time decay costs it less ($7 against $20 if IVs had held).
• The debit spread’s net vega is small, about $0.98 per point, so the crush costs it much less: about $35 at an unchanged price, against about $113 for the outright call. The price of that protection is a capped gain: $260 at $108 against $511 for the call.
• Net vega can point the wrong way. The calendar’s net vega is +$6.58 per point, which says a uniform fall in IV should hurt it. Under the event crush it gains $57 at an unchanged price, against $13 if all IVs had held, because the short front-month leg loses more value than the long back-month leg. It still loses on large moves.

These are European-style model values. Because XYZ pays no dividend, early exercise does not change the call values, but American-style short options still carry assignment risk in practice.

Volatility skew and smile in scenarios

Black–Scholes–Merton implies one volatility for every strike at a given expiration; Rubinstein (1994) described this as a minimal prediction of the formula. Market IVs vary by strike instead. Since the 1987 crash, S&P 500 index options have shown a persistent skew, with higher IVs at lower strikes (Derman, 1999). When IVs rise toward both wings, the pattern is called a smile. Compared with a flat-volatility lognormal benchmark, a downward skew is usually read as pricing more risk-neutral probability of large declines, and a smile as pricing more probability of large moves in both directions. These are interpretations of risk-neutral prices, not forecasts. The full implied distribution depends on option prices across all strikes, and methods such as Rubinstein’s implied binomial trees infer it from those prices.

For scenario analysis, the practical question is what each option’s IV does when the stock moves. Two of the simple rules Derman examined are sticky strike, where each strike keeps its IV, and sticky moneyness (often called sticky delta), where IV depends on the strike’s distance from the stock price and therefore moves with the stock.

Hypothetical 45-day skew: 30% at the 90 strike, 27.5% at 95, 25% at 100, 22.5% at 105, and 20% at 110. XYZ drops instantly from $100 to $95. Profit or loss per contract:

These are European-style values; with American-style exercise, the put results would be about $2 higher.

The same price move produces results about $70 per contract apart depending on the volatility rule alone. For equity options, IV tends to fall when the stock rises and rise when it falls (Hull, Fields Institute talk abstract), the direction of the third row. Derman found, however, that different rules described S&P 500 behavior in different periods, so no single rule is safe to assume.

Key takeaways

• Implied volatility is the volatility that makes a model reproduce an option’s price. It depends on the model, the inputs, and the time convention, and it differs by strike and expiration.
• IV must be solved numerically, commonly by Newton–Raphson with vega as the slope. Bid-ask spreads translate into wide IV ranges for cheap out-of-the-money options, and an American option trading at its exercise value has no well-defined IV.
• An annualized IV scales with the square root of time. A one-standard-deviation range is asymmetric in price terms and describes a risk-neutral distribution, not a forecast.
• IV and realized volatility measure different things over different windows. In the index-option studies cited, option prices implied more variance than was later realized on average, consistent with a negative variance risk premium, and the evidence on IV as a forecast is mixed.
• IV crush is largely mechanical. A scheduled event adds a fixed block of variance that dominates short-dated IV and disappears after the announcement, so front-month IV falls far more than back-month IV.
• When price, time, and IV change together, attributing profit or loss to “the crush” depends on the order of calculation. State the convention.
• Net vega assumes a parallel IV shift. Event crushes and skew moves are not parallel, which is why positions with similar vega can respond very differently.
• IV describes prices, not outcomes. It cannot say what volatility will be realized or what a position will earn, every scenario rests on an assumption about how IV evolves, and model values leave out bid-ask spreads, early exercise, and assignment.

Primary sources

• Black, F., and Scholes, M. (1973). The Pricing of Options and Corporate Liabilities. Journal of Political Economy, 81(3), 637–654.
• Merton, R. C. (1973). Theory of Rational Option Pricing. Bell Journal of Economics and Management Science, 4(1), 141–183.
• Latané, H. A., and Rendleman, R. J. (1976). Standard Deviations of Stock Price Ratios Implied in Option Prices. Journal of Finance, 31(2), 369–381.
• Cox, J. C., Ross, S. A., and Rubinstein, M. (1979). Option Pricing: A Simplified Approach. Journal of Financial Economics, 7(3), 229–263.
• Patell, J. M., and Wolfson, M. A. (1979). Anticipated Information Releases Reflected in Call Option Prices. Journal of Accounting and Economics, 1(2), 117–140.
• Manaster, S., and Koehler, G. (1982). The Calculation of Implied Variances from the Black-Scholes Model: A Note. Journal of Finance, 37(1), 227–230.
• Brenner, M., and Subrahmanyam, M. G. (1988). A Simple Formula to Compute the Implied Standard Deviation. Financial Analysts Journal, 44(5), 80–83.
• Canina, L., and Figlewski, S. (1993). The Informational Content of Implied Volatility. Review of Financial Studies, 6(3), 659–681.
• Rubinstein, M. (1994). Implied Binomial Trees. Journal of Finance, 49(3), 771–818.
• Christensen, B. J., and Prabhala, N. R. (1998). The Relation Between Implied and Realized Volatility. Journal of Financial Economics, 50(2), 125–150.
• Derman, E. (1999). Regimes of Volatility. Risk, April 1999.
• Bakshi, G., and Kapadia, N. (2003). Delta-Hedged Gains and the Negative Market Volatility Risk Premium. Review of Financial Studies, 16(2), 527–566.
• Carr, P., and Wu, L. (2009). Variance Risk Premiums. Review of Financial Studies, 22(3), 1311–1341.
• Hull, J. Optimal Delta Hedging, Fields Institute talk abstract (2015).
• Dubinsky, A., Johannes, M., Kaeck, A., and Seeger, N. J. (2019). Option Pricing of Earnings Announcement Risks. Review of Financial Studies, 32(2), 646–687.
• Cboe: Volatility Index Mathematics Methodology · Volatility Index Methodology: Cboe Volatility Index
• Options Industry Council: Options Exercise FAQ