Implied Volatility

Implied volatility is the level of future price movement that an option's current market price implies, given everything else about the contract. It is found by taking the price at which the option is actually trading, together with the strike, the time to expiration and the other inputs, and solving a pricing model for the one quantity left over. The SEC's staff accounting guidance states both halves of that: implied volatility "is the volatility assumption inherent in the market prices of a company's traded options or other financial instruments that have option-like features," and it "is derived by entering the market price of the traded financial instrument, along with assumptions specific to the financial options being valued, into a model based on a constant volatility estimate (e.g., the Black-Scholes-Merton closed-form model) and solving for the unknown assumption of volatility."

The distinction that matters is the direction of the arithmetic. Every other input to an option's price is observable: the strike is printed on the contract, the expiration is a date, the underlying's price is quoted. Implied volatility is the residual, the number that makes the model agree with the market. So it is a reading of what participants are collectively willing to pay, expressed in the units of a volatility statistic.

Volatility as a statistic, and the general contrast between the realized and implied families of measure, is covered on our page for volatility, which is where the realized side belongs. The best-known implied measure, the index built from S&P 500 option prices, has its own page.

The units are the first thing to get straight, because the number looks like a return and is not one. Implied volatility is quoted as an annualized percentage: a stock trading with 30 percent implied volatility is not expected to return 30 percent, and is not expected to move 30 percent. The figure describes the width of a distribution over a year under the model it was solved out of. Because that model scales volatility with the square root of elapsed time, the implied movement over a shorter horizon is smaller than the headline number in a specific way, which the worked example below sets out. Note also that the convention differs: some sources annualize over calendar days and some over trading days, so two quoted figures for the same contract can differ for no reason other than the denominator.

A high reading is a statement about magnitude, and treating it as a directional signal is the commonest misuse. An option's price rises when either a large fall or a large rise becomes more likely, because both would make some contract worth exercising. So implied volatility rising before a scheduled event says the market expects a wide range of outcomes, not that it expects a bad one. This is the same property that makes the volatility statistic itself symmetric, set out on our page for volatility.

What a high reading means depends entirely on which side of the contract you are on, and the two sides are exact opposites. For a buyer, high implied volatility means the contract is expensive: more of the premium is the price of expected movement, and more movement is needed to make the position pay. For a writer, high implied volatility means the premium received is larger, which is compensation for a risk the market has just marked up. Neither is an advantage on its own, because the price is high or low for a reason that applies to both sides equally.

The most expensive lesson in this subject is that a correct directional call can still lose money when implied volatility falls. The SEC's own list of what determines an option premium has three items: the underlying stock price relative to the strike, the time remaining, and "the price volatility of the underlying stock." A buyer who pays a premium inflated by high expected volatility ahead of a scheduled announcement, and who is right about the direction, can still find the contract worth less afterwards, because the uncertainty the premium was paying for has been resolved and that component of the price has gone. Traders call this an implied volatility crush; the mechanism is simply the third item on the SEC's list falling while the first moved less than the premium needed it to.

It is model-dependent, which is a real limitation rather than a technicality. The SEC's guidance is explicit that the derivation runs through a model "based on a constant volatility estimate," naming Black-Scholes-Merton as the example. Different models, and different assumptions about dividends and interest rates, return different implied volatilities from the same observed price. So the figure is a translation of a price into a model's vocabulary, and it inherits whatever that model gets wrong.

The constant-volatility assumption is visibly violated by the market itself, and the SEC's own guidance implies as much in passing. When SEC staff set out what a company should weigh in relying on implied volatility from its traded options, one of the listed considerations is "the similarity of the exercise prices of the traded options to the exercise price of the newly-granted share options," and another is the similarity of their terms. Those considerations only matter if implied volatility differs across strikes and across expirations on the same underlying, which it routinely does. The pattern of that variation across strikes is what traders call the volatility skew or smile. The practical consequence for a reader is that quoting a single implied volatility for a stock is a summary of a surface, not a measurement of a thing.

There is a persistent gap between what options imply and what markets subsequently deliver, and the index sponsor publishes it. Cboe states that "over long periods, index options have tended to price in slightly more uncertainty than the market ultimately realizes," and specifically that "the expected volatility implied by SPX option prices tends to trade at a premium relative to subsequent realized volatility in the S&P 500 Index." That is a statement about a long-run average on one index, not a rule that holds in any given month, and the periods in which it fails are exactly the periods in which a writer of options loses a great deal quickly. Both halves of that sentence have to travel together.

Take an option's observed market price, together with the strike, the time to expiration, the underlying's price and the model's other inputs, and solve the pricing model for volatility. The answer is the implied volatility of that contract. Repeat across strikes and expirations and the answers differ, which is why quoting one figure for a stock is a summary.

A hypothetical illustration of what the number says about expected movement, under the constant-volatility model it is solved out of. Suppose a stock trades at $100.00 and options on it carry an implied volatility of 30 percent.

Over one year, that figure describes a one-standard-deviation range of about plus or minus $30.00, because 30 percent of $100.00 is $30.00. Under the model, volatility scales with the square root of elapsed time, so over a 30-day window the same figure implies a much smaller range: the square root of the fraction 30/365 is about 0.287, and 30 percent times 0.287 is about 8.6 percent, or roughly plus or minus $8.60 on a $100.00 stock.

Two cautions travel with those numbers. First, "one standard deviation" is a property of the model's assumed distribution, not a probability the market has promised, and real return distributions have fatter tails than the model assumes, so large moves happen more often than it implies. Second, the annualization convention matters: using 252 trading days instead of 365 calendar days over the same window gives a different figure from the same input. All figures are illustrative.

What implied volatility is good for

• It converts an option's price into a comparable number, so contracts on different underlyings at different strikes can be judged as expensive or cheap relative to each other.
• It is forward-looking, which no measure computed from past prices can be.
• It reprices continuously, so it registers a change in expectations the moment market participants act on it rather than after the fact.
• Because it is derived from a price rather than assumed, it can be compared against what the underlying subsequently does, which is how the gap between expected and realized movement gets measured at all.

Where it misleads

• It says nothing about direction. A high reading is consistent with a large rise and with a large fall, and it is often read as a bearish signal.
• It is model-dependent, so the same market price yields different figures under different assumptions.
• It varies by strike and by expiration on the same underlying, so a single quoted number for a stock is a summary of a surface.
• Options have tended to price in more uncertainty than subsequently arrives, which sounds like a standing edge for writers and is not, because the exceptions are concentrated and severe.
• A buyer who is right about direction can still lose money if the reading falls after a scheduled event resolves, because part of what they paid for was the uncertainty itself.
• It is not a probability. The one-standard-deviation framing comes from the model's assumed distribution, whose tails are thinner than real markets'.