2.5 Implied Volatility — What It Is and Why It Moves

2.5 Implied Volatility — What It Is and Why It Moves

Module 02 · Lesson 2.5 · Estimated read 9 min

Implied volatility is the most misunderstood number on the options chain. Retail traders treat it as a forecast: “IV is high, the market expects volatility.” That description is wrong in a specific and important way. IV is a market-clearing price, not a forecast. It is the level at which sellers of variance and buyers of variance found a deal at this strike, at this expiry, this moment. The forecast is downstream of the price, not upstream. This lesson separates IV from realized volatility, walks the term structure of IV, explains why earnings produces the recurring “IV crush,” and gives you the IV rank vs IV percentile distinction that decides whether the IV print on your screen is high or low.

1. IV vs realized volatility

Realized volatility (RV) is what actually happened: the standard deviation of recent returns, annualized. RV looks backward. Implied volatility (IV) is what the market is pricing forward: the back-solved sigma that reconciles current option prices with Black-Scholes. IV looks forward but it is a market-clearing level, not a forecast in the statistical sense.

The relationship between IV and RV is structural: across long samples, IV runs above RV by 2-4 vol points in equity index land. That gap is the variance risk premium — the extra premium that variance buyers (option longs) pay to variance sellers (option shorts) for absorbing tail risk. Selling variance has a positive expectancy in expectation because the expected payoff from having paid IV is lower than the actual realized variance the seller absorbs. The premium compensates for the seller’s asymmetric risk: bounded gain when calm holds, unbounded loss when it does not.

The IV/RV ratio is the cleanest read on whether the market is pricing variance richly or cheaply at the moment. A 30-day SPY IV of 14 against trailing 30-day RV of 9 puts the IV/RV ratio at 1.56 — option sellers are getting paid generously over actual realized variance. A 30-day IV of 22 against trailing RV of 28 means the market is pricing less variance than just realized — the seller is no longer getting a premium and may be on the wrong side. Reading the ratio matters more than reading either number alone.

2. IV term structure refresher

IV is not a single number; it is a function of strike and tenor. The term structure of IV plots ATM IV against time-to-expiry. In normal markets, the curve slopes up (contango): near-dated IV is below far-dated IV because longer horizons accumulate more uncertainty. In stress markets, the curve inverts (backwardation): near-dated IV jumps above far-dated, signalling that the market expects whatever is happening to resolve before the longer-dated options expire.

The conventional summary of the SPX term structure is the VIX-to-VIX3M ratio. VIX measures 30-day expected volatility on SPX; VIX3M measures 90-day expected volatility. VIX below VIX3M is contango (calm regime); VIX above VIX3M is backwardation (stress regime). Module 03 covers this in depth, but it shows up here because IV mechanics cannot be understood without the term-structure dimension.

Practically, a stock with 30-day IV at 24 and 90-day IV at 26 is in normal contango. A stock with 30-day IV at 38 and 90-day IV at 28 is in steep backwardation — usually because of a near-term catalyst (earnings, FDA decision, court ruling) that the 30-day options are pricing through but the 90-day options are not. Earnings is the most common driver of name-level term-structure inversion.

3. Earnings IV ramp and crush

The recurring earnings cycle in IV looks identical across thousands of single names. Two to four weeks before the print, near-dated IV starts to ramp as buyers of pre-event optionality bid the strikes. The ramp accelerates in the final week and especially in the final two sessions. The day after the print, IV crashes — the “IV crush.” The crush is mechanical, not a market reaction to news content.

The mechanic: the earnings event is a known jump risk. The 30-day IV before the event prices in both the regular diffusion variance and the discrete jump expected from the print. After the print resolves, the jump-risk component is removed from the remaining tenor, and IV collapses to roughly the implied forward variance for the post-event period — closer to long-run RV than to pre-event IV.

A worked example: NVDA before its November 2024 earnings print. The day before earnings, the front-week call ATM IV was around 220, the front-week put ATM IV around 240. The day after, both crashed to around 70-75 within the first ten minutes of the cash session. The stock moved less than 5% on the print — smaller than the implied move had priced — and the IV collapse alone destroyed the value of long ATM straddles bought into the print regardless of direction. Buyers who held single legs unhedged lost on the IV crush even when their directional call was correct.

The structural read for traders: the IV ramp is tradable from the long side only if the post-event move exceeds the implied move. Buying ATM straddles into earnings is a coin-flip on whether the realized jump exceeds the priced jump. Across a large sample of single-name earnings, the realized move averages roughly 80-90% of the implied move — meaning straddle buyers lose modestly in expectation. Selling premium into earnings is positive-expectancy on average, but the tail when a stock prints a multi-sigma move can wipe out months of small wins.

4. Why market-makers price IV from supply and demand

The most important sentence in this lesson: market-makers do not forecast volatility. They price options to clear flow. IV is the back-solved level where buy-side demand and sell-side supply intersected at the strike on the chain. When demand for puts exceeds supply, put IV rises. When demand for calls exceeds supply, call IV rises. The IV print is downstream of the trading interest.

This is why IV moves can lead spot. The hedging desk that sees a buyer lifting 25-delta puts in size raises the IV they will quote on the next 25-delta put trade. The next buyer pays a higher price. Spot has not moved. The surface has re-priced because of order flow, not because the market has “learned something new.” The information is in the flow, and IV is the screen that displays it.

The corollary: there is no central voice deciding what IV “should” be. There is a competitive market of dealers who quote tight enough to attract flow but wide enough to cover their hedging risk. The IV you see is the equilibrium of those quotes. When dealers’ inventory tilts long-vol, they lower offers to lay it off. When they tilt short-vol, they raise bids to source it. The screen IV moves in response to those inventory dynamics.

This framing changes what the IV print means. A high IV print does not mean “the market expects a big move.” It means “buyers are willing to pay this much for variance, sellers are willing to provide it at this level, and the clearing equilibrium is here.” The forecast is built by the participants, not by some omniscient auctioneer. Reading IV as a market price rather than a forecast is more honest about what the number represents.

5. IV rank vs IV percentile

Two metrics dominate practitioner usage when answering “is current IV high or low.” They are not the same metric and the distinction matters.

IV rank normalizes current IV against the high and low of the past 52 weeks. The formula:

An IV rank of 50 means current IV sits exactly halfway between the year’s low and high. A rank of 90 means IV is near its 52-week high. A rank of 10 means IV is near its 52-week low.

IV percentile measures the share of trading days over the past 52 weeks where IV closed below the current level. The formula:

An IV percentile of 80 means IV closed below current on 80% of the past year’s sessions. An IV percentile of 20 means IV closed below current on only 20% of sessions.

The two metrics diverge meaningfully when the IV distribution is skewed. After a big spike followed by a long calm period, IV rank can stay near its low (because the spike anchored the high) while IV percentile is near 50 (because the calm period dominated the distribution). IV percentile is closer to the answer to “is today’s level unusual?”; IV rank is closer to “where in the year-to-year envelope are we?”

Both metrics fail in regime shifts. A stock that traded at 30 IV all year and now trades at 60 has an IV rank near 100 and an IV percentile near 100. Whether that is “high” depends on whether the new regime is a temporary shock or a permanent shift — and the metrics cannot distinguish. Traders use them as starting points, not endings, and combine them with term-structure shape and skew shape to read whether the IV print is genuinely elevated or just elevated relative to a quiet anchor.

Key takeaways

• IV is a market-clearing price, not a forecast. The number is what cleared the auction at this strike, this tenor, this moment. The forecast is built by participants downstream.
• IV runs above RV in expectation. The 2-4 vol point gap is the variance risk premium that compensates option sellers for absorbing tail risk.
• Earnings IV crush is mechanical. The known jump risk is priced into pre-event IV; once the event resolves, IV collapses to forward variance, regardless of whether the actual move was large or small.
• Market-makers do not forecast vol; they price flow. IV moves in response to order-flow imbalances and dealer-inventory dynamics. The print is the equilibrium, not the forecast.
• IV rank vs IV percentile. Rank is “where in the 52-week envelope”; percentile is “how many days closed below current.” The two diverge when the IV distribution is skewed, and percentile is usually the more honest read.