Volatility Skew and the Options Smile

Volatility Skew and the Options Smile

Black-Scholes assumes a single volatility per expiry. Reality assigns each strike a different one. The shape that emerges — the smile in FX, the skew in equities — encodes the market’s expectations about the asymmetry of returns. Learning to read it is learning what crash insurance costs and who is bidding for it.

3.4.1From Smile to Skew

The original Black-Scholes (1973) assumed log-normal returns and a single constant volatility. If that were true, plotting implied vol against strike for a fixed expiry would produce a flat line. Pre-1987, equity markets did roughly that. Then 19 October 1987 happened — Black Monday, S&P 500 down 20.5% in a single session — and the market never went back.

3.4.1.1The smile (FX, commodities)

In FX and many commodities, both far OTM puts and far OTM calls trade at higher implied vol than ATM strikes. Plotted, this looks like a smile. The reason is symmetry of risk — traders fear large moves in either direction equally.

3.4.1.2The skew (equities)

In equity indices, the smile becomes a skew. OTM puts trade at materially higher IV than ATM. OTM calls trade at lower IV than ATM. The curve slopes downward from left to right. This is the equity put skew, and it has been a permanent feature of SPX options since 1987.

3.4.2The Equity Put Skew

Why do equities skew while FX smiles? Because equity returns are asymmetric: they crash but they do not melt up the same way. Markets fall faster than they rise — the average bear market sheds 35% over 11 months; the average bull adds 120% over 5 years. Crashes are violent and concentrated. Rallies are slow and diffuse.

3.4.2.1Who pays for downside protection

Pension funds, insurance companies, and family offices structurally bid puts. They have to deliver returns and meet liabilities; tail risk is asymmetric for them. They pay up for OTM puts every month, every year. That structural demand is what makes the skew permanent.

3.4.2.2Who supplies it

Volatility-arb desks, dealers, retail premium-sellers — anyone willing to be short tail risk in exchange for the steepness premium. The skew premium is real but the path can be brutal: the 2018 vol-mageddon, March 2020, August 2024 each obliterated names that had been collecting steady skew premium for years.

3.4.3Risk Reversals

The cleanest single-number measure of skew is the 25-delta risk reversal: the IV of the 25-delta call minus the IV of the 25-delta put for a given expiry.

3.4.3.1Sign convention

• RR < 0: Put skew — puts more expensive than calls. Equity-index default state.
• RR > 0: Call skew — calls more expensive than puts. Common in single-name takeover targets, commodity squeezes.
• RR ≈ 0: Symmetric — rare in equities, normal in FX.

SPX 25-delta 30-day RR has historically averaged about -3.5 vol points. In calm regimes it sits at -2 to -3. In stressed regimes it can reach -8 to -12. As of 5 May 2026: -3.1 — squarely in the calm zone, consistent with VIX 17.32 and a normal contango regime.

3.4.3.2Skew steepness vs vol level

An important nuance: skew steepness can rise even when vol level falls. In 2017, VIX hit historic lows but the put skew was simultaneously near record steeps — the market was calm but paying through the nose for tail protection. That divergence — cheap ATM, expensive tails — is one of the cleanest complacency-with-fear signals on the tape.

3.4.4Reading Skew Steepness

3.4.4.1The skew percentile

Better than absolute level: skew percentile vs trailing 252-day distribution. Today’s -3.1 sits at roughly the 42nd percentile — squarely middle-of-the-distribution, no signal in either direction.

“The skew is the price of fear. When it is cheap, fear is unfashionable — which is when you should be buying it.”

3.4.5Case Study: Crash Skew

3.4.5.1February 2020 — the COVID skew spike

Through January 2020, SPX 25d 30d RR averaged -3.2. On 24 February 2020, as the first credible COVID risk-off began, the RR widened to -5.8 in a single session — a 2.6 vol-point move when the average daily change is roughly 0.3. Spot SPX was only down 3.4% that day. By the 28th, before the worst of the drawdown, RR had blown out to -9.5.

The lesson: skew typically moves before spot moves enough to justify it. By the time the index was down 10%, RR was already at extreme levels and crash positioning was visible across the surface.

3.4.5.2August 2024 — yen carry skew

The 2 August 2024 close — the Friday before the Monday spike — saw 25d 30d RR jump from -3.5 to -7.8. Spot SPX was down only 1.8% that day. The skew widening foretold the Monday gap to a degree that price action alone did not.

3.4.6Trading Skew

3.4.6.1Sell rich puts in calm regimes

When skew is at the 70th+ percentile and VIX is under 20, OTM put-selling earns the steepness premium. Best as cash-secured puts on names you would want to own anyway. Sized so a 2-sigma down day does not threaten the book.

3.4.6.2Buy puts when skew is cheap

When skew is at the 30th percentile or lower, OTM puts are unusually cheap relative to ATM vol. Long-vol setups via OTM put spreads provide tail hedge with minimum carry cost.

3.4.6.3Risk reversals as directional plays

Selling a 25d put and buying a 25d call (zero-cost or small debit risk reversal) is a leveraged bullish bet that benefits from skew compression on rallies and the underlying drift. Common in single names where call skew is building (NVDA pre-AI-narrative was a classic).

3.4.6.4How SomerQuant uses skew

The Sentiment sub-index of the MPI consumes 25d 30d RR percentile as one of its inputs alongside put/call ratios, AAII positioning, and credit spreads. Extreme percentiles in either direction — below 10th or above 90th — contribute as fear/complacency signals. Today’s 42nd percentile reads as neutral.

3.4.7Common Mistakes

• Confusing skew level with skew percentile. -4 RR is below average in 2020 but a 90th-percentile reading in 2025. Always normalize.
• Buying tail puts at peak skew. When everyone wants protection, you pay the most. Crash skew has historically reverted within weeks.
• Ignoring single-name skew specifics. Biotech, cannabis, meme stocks often carry call skew, not put skew. SPX rules do not transfer.
• Trading skew without monitoring spot vol. Cheap skew with low VIX is a different setup from cheap skew with VIX rising.
• Misreading the term structure of skew. 30-day skew and 90-day skew can disagree. Front-month spike + flat back month is event-driven; both wide is regime stress.