Skew Harvest (Systematic Risk Reversals)

Overview

If you plot implied volatility against strike for equity index options, you do not get a flat line. You get a downward slope: out of the money puts trade at much higher implied volatilities than out of the money calls at the same distance. This is the volatility skew, and it has been a permanent feature of equity index options ever since the crash of 1987 taught everyone that markets fall much faster than they rise.

Two explanations, both true. First, equity returns genuinely are negatively skewed. Big down days really are bigger and more common than big up days. Second, and larger in effect: everyone with a stock portfolio wants downside protection and hardly anyone wants upside speculation in size, so the demand for puts is structurally greater than the demand for calls.

The skew harvesting trade says: the second effect is bigger than the first. The puts are priced for a crash that is worse than the crash that actually tends to arrive. Sell the expensive put, buy the cheap call, pocket the difference. That structure is called a risk reversal.

Strategy logic

• The position: Sell a 25 delta put, buy a 25 delta call. Same expiry, matched deltas. The put brings in more premium than the call costs, because the put's implied volatility is higher. You start with a net credit.
• The bet: You are betting the surface is wrong about how asymmetric the future will be. You want the market to grind up, drift sideways, or fall only gently.
• Delta hedge or not: Unhedged, the position is long delta by construction and behaves a lot like owning stock. If you want a clean bet on skew itself rather than on direction, you must delta hedge, and then the whole P&L is about whether the skew flattens or steepens.
• The essential modification: Buy a further out of the money put to cap the loss. Without that, this is a naked short put position with an extra call bolted on, and the honest maximum loss is the entire distance to zero.

Parameters (knobs)

• Delta of the wings: 25 delta is the market convention. Going further out (10 delta) makes the skew look even richer in implied volatility terms but the premium in cash terms gets thin while the tail risk stays vivid.
• Skew percentile filter: This is the difference between a strategy and a habit. Skew is not constant. It steepens dramatically after a shock and flattens in complacent markets. Selling skew when it is already flat means you are being paid almost nothing to hold a crash. Only put the trade on when skew is genuinely elevated.
• Wing protection: How far out you buy the protective put. Closer means a cheaper worst case and a smaller credit. Take the cheaper worst case.
• Tenor: One to three months. Short-dated skew is noisier and moves faster.

Where it works and where it hurts

It works in the long, boring, upward-drifting stretches that make up most market history. The put you sold expires worthless, the call you bought sometimes pays, and the skew you sold slowly flattens as complacency sets in.

It fails in a way that is worth being precise about, because it has two distinct failure modes:

• The obvious one: The market crashes. Your short put is now deeply in the money and your long call is worthless. You lose. Everyone expects this.
• The one that catches people: The market does not crash at all, but skew steepens. Something spooks the market, demand for puts jumps, and the put you are short reprices sharply higher even though the index has barely moved. You take a mark to market loss on a trade whose thesis has not yet been disproven, and if you are levered or margined, you may be forced out before you find out whether you were right.

That second mode is the one that turns a survivable trade into a fatal one, because it hits you before the crash, not during it.

Backtest design checklist

• You need a real, historical implied volatility surface. Not a model. Not an approximation. The entire signal is the shape of the surface, so if your surface is synthetic, your backtest is measuring nothing.
• Measure skew by delta, not by strike. A fixed strike moves relative to spot as the market moves, so a strike-based skew measure conflates skew changes with market direction.
• Backtest the capped version. The uncapped version's worst outcome is a number so large it invalidates every other statistic in the report.
• Separate delta P&L from skew P&L. Unhedged, most of your return is just equity beta wearing a costume. If you cannot show that the skew leg made money independently, you have not found an edge, you have found leverage.
• Include the steepening scenario, not just the crash scenario. Model what a 5 percent index drop with a large skew steepening does to your marks.

Common failure modes

• Selling skew because it looks rich in volatility points. A 10 delta put always has a much higher implied vol than a 10 delta call. That is not a signal. That is the permanent state of the world. The signal is whether it is rich relative to its own history.
• Running it unhedged and calling it an options strategy. An unhedged risk reversal is mostly a long equity position. If the market goes up you will feel clever and you will have learned nothing.
• No wing. The naked short put is the entire risk of the trade. Not capping it is the single most common way this ends badly.
• Forgetting why the skew exists. It is compensation for real, observed, negative skewness in equity returns. Some of that premium is fair. You are only harvesting the excess above fair, and the excess is a lot smaller than the headline number.

Our notes and suggestions

Skew harvesting is intellectually elegant and it has a genuinely sound economic story behind it. It is also, functionally, a bet against the exact scenario that hurts every other asset you own. It is negatively correlated with your own survival.

If your portfolio already holds equities, and it almost certainly does, then selling skew is doubling down on the thing you are already exposed to. The diversification benefit is not merely zero, it is negative. That is worth sitting with for a moment before you put the trade on.

Run it capped, run it only when skew is in the top of its historical range, run it small, and understand that a run of quiet profitable months is not evidence that the risk went away. What would change our mind: implied skew falling persistently below realized skew, meaning the market has stopped overpaying for downside protection, at which point there is nothing left to harvest and only the crash exposure remains.