Skew Trading and Risk Reversals
Skew trading expresses a view on the relative prices of downside and upside options rather than merely on the level of implied volatility. The canonical equity expression is a risk reversal: buy an out-of-the-money put and sell an out-of-the-money call, usually at equal absolute delta. It is deceptively familiar. A risk reversal carries material delta, changing vega, vanna, and jump exposure, so it is neither a pure skew instrument nor a cheap substitute for a forward.
Define the quote before the trade
Markets often quote a 25-delta risk reversal as:
RR_25 = IV(call, delta=+0.25) - IV(put, delta=-0.25)
Equity-index skew is usually negative under this convention: downside puts trade at higher implied volatility, making RR_25 < 0. FX conventions can differ in premium-adjusted delta, delivery date, and whether the sign is put minus call. Record the convention with each observation; otherwise an apparent signal can simply be a sign or delta-definition error.
| Structure | Long position | Primary exposure | Common use |
|---|---|---|---|
| Long put risk reversal | long put, short call | more negative skew | crash insurance or skew widening |
| Long call risk reversal | long call, short put | less negative / positive skew | recovery or upside-skew view |
| Put spread | long lower put, short higher put | local downside curvature | bounded downside view |
| Butterfly | wing options versus centre | smile curvature | convexity relative value |
Risk reversals should be interpreted beside ATM level and butterfly curvature. A put can become richer because all vol rises, because downside skew steepens, or because the smile becomes more curved. Only the second is directly captured by a matched-delta risk-reversal change.
Delta conventions change the object
Strike-based comparisons are unstable when spot moves. Delta-based points better preserve moneyness, but delta itself depends on volatility and the model. Under Black-Scholes,
Delta_call = exp(-qT) * N(d1)
d1 = [log(F/K) + 0.5*sigma^2*T] / (sigma*sqrt(T))
An interpolator must solve for the strike whose market-convention delta equals the target. Holding yesterday's strike fixed after a selloff mixes a change in skew with a change in moneyness. Holding delta fixed is cleaner for surface monitoring, while fixed strike may be appropriate for the actual option held. Report both.
def risk_reversal(surface, forward, expiry, abs_delta=0.25):
k_call = surface.strike_for_delta(forward, expiry, abs_delta, "call")
k_put = surface.strike_for_delta(forward, expiry, -abs_delta, "put")
call_iv = surface.iv(k_call, expiry)
put_iv = surface.iv(k_put, expiry)
return call_iv - put_iv, (k_call, k_put)
The function assumes a surface with consistent forwards, discounting, and deltas. Production code should reject extrapolated deltas, retain bid/ask inputs, and calculate sensitivities on the same pricing convention used by execution.
Trade construction and hedging
Choose strikes first, then size legs. Equal option counts approximate a quoted risk reversal, but equal vegas better isolates a relative-IV move. Neither produces zero delta. Delta hedge the initial package in stock, futures, or swaps, subject to funding and borrow. Rehedging converts the package into a path-dependent gamma strategy; too frequent hedging can turn a small skew edge into transaction costs.
| Hedge decision | Benefit | Cost |
|---|---|---|
| No delta hedge | preserves intended directional view | large spot P&L obscures skew P&L |
| Initial hedge only | separates entry exposure | delta returns as spot and IV move |
| Band hedge | controls tail delta | threshold and slippage model matter |
| Continuous model hedge | theoretical benchmark | impossible across gaps |
For risk reporting, use the options Greeks plus surface factors. Parallel vega says little about a risk reversal: a put and call can have offsetting vegas while the position remains highly exposed to a skew twist. Vanna is especially important in equities because spot declines and IV rises together; the long put gains from both its direct price move and the surface response.
What drives skew?
The leverage effect, demand for crash insurance, dealer hedging flows, and index composition all contribute to negative equity skew. Single names are often governed by event risk and borrow. FX skew can reflect balance-of-payments hedging and crash direction. These are risk-premium explanations, not mechanical forecasts: a persistent negative skew is compatible with puts still being expensive on average.
A useful cross-sectional signal compares current RR_25 with its own history after conditioning on ATM IV, maturity, and spot drawdown. A raw z-score is fragile because skew distributions are asymmetric and jumpy. Estimate robust median/MAD scores, apply liquidity filters, and treat earnings, central-bank decisions, and dividend dates as separate regimes.
Failure modes
Selling downside skew can earn premium repeatedly and then lose multiples of accumulated carry during a gap. A short put risk reversal is short a low-strike put precisely when liquidity disappears and correlation between spot and vol becomes most adverse. Do not use historical daily delta-hedged P&L alone; inject discrete gaps, volatility-surface shocks, borrow recalls, and widened bid-ask spreads.
Calibration also matters. A SABR volatility model or spline can yield a smooth smile while moving the quoted 25-delta point. Monitor raw executable quotes separately from model marks, and do not trade a one-vol-point signal inside a two-vol-point combined spread.
Key takeaways
- A risk reversal is a relative skew trade with meaningful directional and jump exposure.
- Define delta, sign, premium convention, forward, and maturity before comparing quotes.
- Separate parallel-vol, skew, and curvature changes in attribution.
- Delta hedging manages direction but creates path-dependent gamma and execution costs.
- Negative skew is a compensation for crash risk, not evidence that selling it is automatically profitable.