I was thinking about creating a pseudo-butterfly of sorts by selling an atm straddle in an etf with higher IV and buying a strangle in another etf with lower IV and lower cost, at a 1:2 ratio (1 short per 2 longs). The two etfs are highly correlated (most component stocks are the same, correlation in the past year has never been below 75%). Ideas, thoughts? How to evaluate profit targets / breakeven points? Risk elements?
They're highly correlated... but yet they have significantly different IVs? That doesn't sound especially correlated to me. Since your broker is probably unlikely to agree that this constitutes a fly, you're now going to pay the full monty in buying power reduction - i.e., have to put up the margin for the short straddle plus the debit for your pair of strangles - and all you'll end up with is the equivalent of a fly with notionally cheaper wings (but possibly not, if they drift toward actual correlation over the tenor of this creature's existence.) I get the basic premise - cheap wings - but I question whether they're actually cheap in terms of the vol you'll be buying in light of that theoretical correlation. Given that, I don't see an actual benefit to doing this - and I do see a number of detriments. It would also be interesting to see how you visualize the profit path for this thing, given that possible 25% disparity.
Gingerbread person mixing with a sailor of the high seas, does not seem a good combo. Gingerbread is water-soluble.
Hello there, Since in my particular market I'm getting a haircut no matter if I'm trading bounded or unbounded risk, I'm not that much concerned about the reduction in buying power, and I can get the whole structure up for a slight credit (actually didn't resist the temptation so I just put on 1 lot to see what happens). The big issue is exactly the one in the second part of your post, the profit path is a big question mark.. no idea how to analyze it nor to determine possible exit points.
Your view appears to be that the IV of ETF A (the high vol one) is too high relative to the IV of ETF B (the low vol one). This may be the case, but you haven't shown any evidence. The intuition that highly correlated ETF's should trade at similar vol is incorrect. Consider an ETF and its 2x analog. Cor will be ~100% but the 2x ETF will trade at twice the vol. However the intuition that cor and vol are mathematically related is correct, and can be used to spot mispricings and quasi-arbs. Draw a triangle. Label one vertex ETF A, one vertex ETF B, and the third USD. The angle at vertex USD should be arccos(.75). The length of side A-USD should be vol of A. The length of side B-USD should be vol B. The length of side A-B, and the angles at vertices A and B can be determined via the law of cosines. The derived length of side A-B is the implied vol of A expressed in units of B. The cosines of the derived angles at vetices A and B are the implied correlations of A:USD and B:USD respectively. Check these implied figures with recent A/B vol and A:USD and B:USD correlations. If the any of the recent realized vol and cors are way off the triangle derived numbers you probably have a trade. Wystrup explains triangle calcs succinctly at this link: https://staff.fnwi.uva.nl/p.j.c.spreij/winterschool/3greeks.pdf Quoting Wystrup: "Computing correlation coefficients based on volatilities as above has a striking implication. Any ... correlation risk can be transformed into a volatility risk"