Okay, this is a good one. Let's talk about volatility risk premium, volatility term structure, volatility spikes, and "the obvious thing often isn't right" (1/?)
https://twitter.com/JessicaNutt96/status/1877829000535454146
Among casual watchers of volatility markets, the most common sentiment is "sell volatility when it's high" or "sell volatility when it spikes". This is oversimplified at best and dangerous and wrong at worst
Fact that you should confirm for yourself with backtests if you are at all serious about trading: there is no reliable empirical relationship between the absolute level of equity index implied volatility and the PNL of buying or selling it. None at all.
Obviously you could come up with extremely over-fit rules that seem to work in-sample, but just make a scatterplot of initial level of implied volatility and subsequent PNL of selling straddles or strangles or VIX futures and draw a line through the cloud of dots, no relationship
When you are trading volatility, you have to think about volatility risk premium, how much is priced in, how you can measure that
One thing you can do is look for where there are persistent differences in realized volatility as compared to implied volatility. (Don't compare S&P realized volatility to VIX, VIX is a variance swap level not an at the money vol, look for my tweets about that)
But as Jessica suggests, another important thing you can look at is volatility term structure. When trading VIX futures this is the single most important thing to pay attention to.
Volatility term structure means the shape of the futures curve, the level of implied volatility by maturity. Is it upward sloping (contango) or backward sloping (backwardation)? How steep is it?
If you buy a second month VIX future, over time all else equal it will roll down the term structure towards where the front month future is. This is the same as in commodities, fixed income and currencies: it is *carry*. Heard of FX carry trades? It's the same thing.
When there is a lot of risk premium in the volatility term structure, when there are a lot of buyers of VIX futures and ETNs, it steepens out the term structure and makes it more expensive to hold a long position in VIX futures
Vice versa, when there is a lot of selling of VIX futures and betting on lower volatility in the future, it flattens or inverts the VIX term structure, making the cost of carry of holding a short position negative (you pay to be short volatility)
The slope of the term structure is empirically a much better signal for how a volatility trade will perform, especially on a market neutral basis (e.g. if you think about a long volatility position beta hedged with long equity). Steeper term structure --> worse long vol returns
Counterintuitively to many non-specialists dabbling in volatility, a volatility spike that brings about steep backwardation in the volatility term structure can be the worst possible time to short volatility
People think "well, volatility spiked, and it's mean reverting, it will eventually go down." Yes, of course, the market knows that and is aggressively pricing in mean reversion. If you short VIX futures, you're betting that volatility mean reverts faster than the market is pricing
And maybe it will! But the market is not dumb, and volatility spikes often persist longer than you think or get worse, and short volatility positions often get crowded and have to blow out before things get resolved
Late February / early March 2020 was a great example of that. VIX hit 40, everyone piled in to short March VIX futures, creating a 14 point differential between VIX spot and the March futures. Volatility had to collapse in record speed for you to make money on a short 
Owning a long position paid you a tremendous carry, in order to be long volatility during a time when the world was very uncertain. And then eventually those March futures hit 80+. Selling them when VIX hit 40 was the one of the worst volatility trades of all time
Allianz Structured Alpha shorted a ton of the March VIX calls by the way, in a double-or-nothing effort to make back its losses on short S&P option structures where they lied about buying back the downside tail. That's one of the flows that made the March futures so cheap
There is a strong relationship on average between the level of implied volatility and the shape of the term structure; higher equity index implied volatility tends to mean a flatter or inverted term structure, low volatility tends to mean steeper
This is also true (or even more so) for recent changes in implied volatility: recent spikes in implied volatility tend to flatten the term structure
However, this is only a correlation, and there is lots of noise around it. There are persistent regimes with relatively high or relatively low risk premium, arising from large structural patterns in volatility buying or selling flows
Implied volatility term structures can have extremely different degrees of steepness and base level, even when in contango. Note the blue line from 2012, high and very steep (very high risk premium), versus purple in early 2018, low and very flat (~negative risk premium)
2012 were the glory days of short VIX trades, after retail investors and RIAs started piling into VXX and TVIX, TVIX suspended creation/redemption, you could just buy put options on VIX or VXX and print money with abandon, 4 points of term structure between VIX and UX1
In contrast, early 2018 was a massive volatility crush fueled by non-specialist volatility traders piling into XIV and similar products. They didn't realize they were short volatility at record levels and getting paid almost nothing for it (no term structure rolldown)
Takeaways:
1) shorting volatility bc its high or buying it bc its low is naive
2) volatility term structure (absolute, and also relative to the level of volatility) is much more interesting as a signal of risk premium than the level of volatility
3) don't trade volatility
1) shorting volatility bc its high or buying it bc its low is naive
2) volatility term structure (absolute, and also relative to the level of volatility) is much more interesting as a signal of risk premium than the level of volatility
3) don't trade volatility
Waiting for "volatility and derivatives in bio. deeply unserious opinion"
Thank you for listening, if you liked your service this morning please do your patriotic duty and buy your copy of Lessons in Birdwatching so that Honey can afford to buy the new Elden Ring coming out this year
• • •
Missing some Tweet in this thread? You can try to
force a refresh
