Harvesting Volatility Risk Premium (VRP) – aka Selling Vol
This paper from Parametric covers three ways to sell equity vol:
- Index options
- Variance swaps
- VIX futures
My Summary and notes from Parametric Portfolio:
Selling Index Options [ex: SPX, RUT, VIX]
He looks at selling straddles and strangles using 20 delta options. 20 delta options are generally far enough away from where the underlying asset is trading that they shouldn’t (but they definitely can) go in the money – while still having enough risk that they hold some value. I.E. 20 delta options are seen as a good risk/reward balance.
Selling straddles and strangles is essentially a bet that implied volatility (at the level you sell) is higher than the realized volatility of the asset. Essentially if you sell the option and the option moves in a major way you could lose. If the stock flat-lines, you should come out ahead.
According to the author selling index options has less volatility than selling variance swaps and/or futures. And if the options are fully collateralize you reduce the chance of margin issues. (You can definitely have a “cash secured” short put, but I don’t think there is such a think as a cash covered short call, because you have essentially unlimited risk).
He calls options “an impure form of trading volatility” because the options are only “indirectly linked to volatility.” The options price is ultimately determined by price of the underlying. I would agree with this in regards to stock options, and some index options however I would say that might not be as true with VIX Index options. (VIX is linked to the value of rolling 30 day options on the S&P500).
Finally he says that “it can be daunting for inexperienced investors to manage an option strategy”. But variance swaps aren’t?
The “swap” in “variance swap” is a misnomer as there are no cash exchanges for the duration of the trade. The variance swap has a fixed strike and expiration. The value of the contract is based on the difference between the “realized variance (volatility squared)” of a given asset (usually an index like the S&P500) and the strike level. The value of the variance swap is related to the vega notional which is a metric used to determine the sensitivity of a derivatives value to the changes in volatility.
Essentially its a more “pure” bet strictly on the volatility of an asset. The seller of the variance swap bets the buyer that volatility will stay under x level. These are called OTC (over the counter) trades and are generally left to banks, hedge funds and the like. The author says these are “the purest” form of trading volatility. They are also easy to leverage and can be tailored to meeting investors specifications.
The downside of these contracts are the “quadratic form of the payoff function”. Which essentially means that the returns are “convex” or upward sloping. A stock for instance has a linear payoff – you make $1 per share if they stock goes up 1 point. When volatility really spikes the variance swap (and all options for that matter) can start with a return of $1/pt then quickly rise to return 4,5,6 and up per point. Put another way its a curved upward sloping (convex) payoff, not a straight line.
Derivatives linked to VIX offer investors a “direct approach to monetize the premium embedded in equity volatility” as the VIX is linked to the S&P500 index options. He alleges that VIX futures are more popular than variance swaps. The VIX futures are a popular form of “tail risk” insurance, but its expensive as they contango of the VIX futures curve makes it expensive to own. I.E. owners of VIX futures lose every month rolling the futures in a “normal” environment. So if you were to sell you might make profit on this roll.
VIX futures are very liquid and are a pretty pure play on volatility. The futures can be very volatile in the short term, distorting pricing. There is a strong equity beta towards VIX futures (linked to the movement of the S&P500). Notional exposure of VIX futures is expressed in vega (as in variance swaps).
I heard someone once say “no one ever lost money backtesting”. I agree, so take the returns below with a grain of salt. In reality you have issues to deal with when actually trading this stuff – margin calls, transaction fees and the like. That being said you can see the significant drawdown on the variance swap. Linking back to the authors note about the “convexity” of the variance swap gain/loss …I would not want to be on the wrong side of that trade.