Vanna is a second-order Greek which measures changes in delta in relation to changes in implied volatility. The mechanics behind it is that when implied volatility increases, then all deltas get closer to 50. This is because an increase in IV means that the implied range becomes wider, and a wider implied range means that strikes need to be further away to have the same deltas. Conversely, if implied volatility decreases, then the implied range becomes smaller and strikes do not need to be as far away to have the same deltas. 

The connection to the typical scenario of IV crush leading to a rally (or at least bullish flows) is when implied volatility drops while the price action is trading in a normal range. We have the general assumption that dealers (market makers) are short puts on a Put Wall. The causal effect of vanna is that dealers buy back shares since they no longer need to short as many shares or futures in order to balance out their short puts in order to stay delta-neutral. The mechanics here is that the decrease in IV causes a decrease in (OTM) out-of-the-money put deltas because the shrinking implied range pushes deltas further OTM. 

Intermediate - Using our Vanna Model to Trade

We share a vanna model each day. This model is available for all option chains on all expirations. A historical view of this information is available with tabs on this model, showing what the model looked like on those past two days. 

Shown here in beige is the forecasted effect of delta as there are moves in the underlying (across the x axis). The higher the beige line is vertically, the more that we model dealers to need to sell shares in order to balance out their directional risk. 

The adjusted effects of vanna are shown with the purple line. This is how much the delta line is modeled to change as a result of changes in IV.  With the underlying shown on the x-axis, where the purple line is shown above the beige line, we model there to be extra selling required from dealers in order to stay balanced. That means a bearish flow contributing to the market because of vanna. And when the purple line is under the beige line, this means that net buying flows are being contributed to the market because of vanna because dealers must buy extra amounts of shares (compared to delta) in order to stay balanced.

What specifically drives our vanna model is our SIV model, which is our projection of IV will change depending on moves in the underlying (along the x axis). This SIV model is more nuanced than a basic assumption that IV always drops when the underlying price increases, and will take into account how sometimes IV can increase during rallies, especially if IV is low enough compared to historical averages.