• C++ Programming for Financial Engineering
    Highly recommended by thousands of MFE students. Covers essential C++ topics with applications to financial engineering. Learn more Join!
    Python for Finance with Intro to Data Science
    Gain practical understanding of Python to read, understand, and write professional Python code for your first day on the job. Learn more Join!
    An Intuition-Based Options Primer for FE
    Ideal for entry level positions interviews and graduate studies, specializing in options trading arbitrage and options valuation models. Learn more Join!

Purest way to trade option skew?

Joined
7/25/10
Messages
862
Points
38
Hey mates, simple question to formulate, not sure if it's trivial to answer.

What's the purest way to trade index skew? On the one hand, it might be tempting to say a delta-hedged risky, but you run into the problem that strikes are fixed, so if underlying moves enough, you might have been correct about where skew will move in log moneyness or delta space but still have negative PnL...


Any ideas?
 
That's a bit short sighted. How is a variance swap constructed? From vanilla options -- the strike price of a var swap reflects market option implied volatility, not time series volatility.
 
Put another way, the replication / hedging strategy is exactly what you need to do to trade skew.
 
That's a bit short sighted. How is a variance swap constructed? From vanilla options -- the strike price of a var swap reflects market option implied volatility, not time series volatility.
Well there's two issues with this:
1) Because of convexity, var swap strike isn't even going to be equal to vol swap strike.
2) Var swaps are not a good way to trade implied vols. Of course, they have vega (which decreases over time!), but it is a really unpure way to trade implied vols, especially as they become more and more seasoned.
 
How about doing a forward starting skew swap, one that only takes observations starting at some point in the future? If you buy, say, a 6 month skew swap which starts fixing in 3 months' time, you get a 6 month implied skew level entry point that you can close out / monetize at any point until it starts taking observations in 3 months. You can also trade a forward starting vanilla risk reversal (which takes strikes based on % moneyness or delta at the end of the forward starting period). In either case, I'm not sure how easy you'll find it to get people to make you a price in real life. Ultimately you're probably stuck with trading vanilla risk reversals and rebalancing when the underlying moves.
 
Well there's two issues with this:
1) Because of convexity, var swap strike isn't even going to be equal to vol swap strike.
2) Var swaps are not a good way to trade implied vols. Of course, they have vega (which decreases over time!), but it is a really unpure way to trade implied vols, especially as they become more and more seasoned.
So if I'm reading you right...
(1) Means you don't want the squared effect? Does this really matter a lot to you? If you can predict future vol, you can predict future var. Also the fair strike var swap won't equal the squared vol swap strike, but again why does this matter?
(2) So you REALLY just want to trade implied vol at a fixed tenor? Like 1 month forward implied vol? There is no cheap way to replicate this, because to create the hedge would involve rolling options every day (or whatever frequency is allowable due to constraints of what's listed).

Verify this for me: "pure" to you means both fixed tenor and independent of underlying price?
 
I think you guys are talking about different things. I understand euroazn's aversion to trading variance swaps in order to trade vol. Yes there is a direct relationship between the two, but in reality there's a ton of risk in selling variance that you don't have when you sell volatility, and in the other direction, there's a premium attached to buying variance that you may not want to have to pay. There is a product called an FVA (forward volatility agreement), which settles into a ATM straddle on some future date. That's the purest way to bet on future implied volatility of vanilla options, since it is directly that. You can also trade a forward starting volatility swap. But in any case, I don't think that's what euroazn is after since he just wants to be able to bet on the future market pricing of implied skew relative to today. For that, I stand by my previous post. You can also look at doing a forward starting variance swap vs a forward starting gamma swap, or the analogous payoffs in linear vol space so that you run into short or long vol, rather than variance, as spot moves.
 
I can't imagine you'd be able to find good two-way prices for some of these 'second generation' vol products like gamma swaps, FVAs, corridor var swaps. It gets even worse once you move outside underlyings such as spx, nky, eurusd, and usdjpy.
 
I can't imagine you'd be able to find good two-way prices for some of these 'second generation' vol products like gamma swaps, FVAs, corridor var swaps. It gets even worse once you move outside underlyings such as spx, nky, eurusd, and usdjpy.

Some are more liquid than others (e.g. gamma swaps are just as replicable with vanillas as variance swaps), some are more liquid depending on what asset class you are in, but of course it's all dependent on getting a decent price. That's why in my original post I said realistically you're probably going to have to deal with trading around and rebalancing vanillas to get the exposure you want.
 
Back
Top