 |
newwurldmn
Registered: Apr 2011
Posts: 2617 |
03-17-12 07:25 AM
Quote from trefoil:
Can't speak for atticus of course, but I thought the proposed structure was brilliant because it's practically like buying a straddle/strangle, if the SPY and VIX act anything like they have historically.
Depends on the ratio of spx delta to the vix. That's always a tricky thing and hard to guage. They will definitely hedge, but to what extent is almost completely unkown except it will be something greater than zero.
|
| |
|
Edit/Delete • Quote • Complain |
trefoil
Registered: Mar 2007
Posts: 3383 |
03-17-12 02:50 PM
Not expecting perfection. When I thought about it, I figured the logic here was that to get back to 23 - which spot has only had one or two brushes with since the year started - would take a sharp decline in the indices. I understand that VIX can fluctuate around in a manner uncorrelated with the indices within a certain range, but by now we're looking at a 50% + spike, and I can't see that happening without a nice decline in SPX, and of course SPY.
|
| |
|
Edit/Delete • Quote • Complain |
newwurldmn
Registered: Apr 2011
Posts: 2617 |
03-17-12 03:43 PM
Quote from trefoil:
Not expecting perfection. When I thought about it, I figured the logic here was that to get back to 23 - which spot has only had one or two brushes with since the year started - would take a sharp decline in the indices. I understand that VIX can fluctuate around in a manner uncorrelated with the indices within a certain range, but by now we're looking at a 50% + spike, and I can't see that happening without a nice decline in SPX, and of course SPY.
Yeah. So the SPY declines 10%, where will the VIX be then? Probably 30+ and implieds on the VIX gone through the roof. So your call will be worth a lot more than $7. How much SPY put do you need to buy for that? And then how does that hedge work if the SPX sells off only 5%? And how well does that work if the SPX rallies 3%?
It's a linear hedge against a convex risk. Very difficult to manage - at least in my experience.
|
| |
|
Edit/Delete • Quote • Complain |
rallymode
Registered: Dec 2005
Posts: 1325 |
03-17-12 04:15 PM
Quote from trefoil:
FYI, I did this but modified it somewhat.
Sold 6 23/28 call spreads on April for the VIX, and bought 2 139/129 put spreads on the SPY. I didn't sell outright 'cuz of the crazed margin I'd have to put up, and I did 23 because it was close to ATM at the time, and I always like to sell/buy at right around ATM if I can manage it these days.
I used to do something similar on a monthly basis for years and it works very nicely though you might wanna go a little farther out on your gamma hedge. If you know how to properly model the var exposure you can easily hedge with cheap otm gamma, provided you don't go unbound on risk. Also, given that the current term structure is more of an outlier than the norm it is not a perfect data point for an experiment unless you weigh it.
|
| |
|
Edit/Delete • Quote • Complain |
trefoil
Registered: Mar 2007
Posts: 3383 |
03-17-12 04:16 PM
Quote from newwurldmn:
Yeah. So the SPY declines 10%, where will the VIX be then? Probably 30+ and implieds on the VIX gone through the roof. So your call will be worth a lot more than $7. How much SPY put do you need to buy for that? And then how does that hedge work if the SPX sells off only 5%? And how well does that work if the SPX rallies 3%?
It's a linear hedge against a convex risk. Very difficult to manage - at least in my experience.
Good point, however I did a call spread, not an outright sell, so max risk on each done is 500-108 = 392 per. 392 * 6 = 2352 max risk overall, against max profit on the put side of 892 * 2 = 1784. Cost of the put spreads was 108 per, so assuming absolute worst case and held to expiration I'd see a loss of 2352 + 216 = 2578. Stinging but not catastrophic.
As you're probably aware, the diff between each strike on the VIX options isn't that high, and expands and contracts only slowly. So, as a practical matter, once you've done a spread the max risk on each spread would only come into play on a truly extreme and rapid move in the VIX, remembering that a rise in the implieds hits both sides of the spread, after all, firstly, and secondly the SPY put spread value would rise as well with a rise in overall implieds that a rise in the VIX is a measure of. Considering how wide I made those put spreads I'd see a nice increase in value on any increase in IV that accompanies a downmove. Overall IV on April puts according to TOS at the moment is only 15.4%, so there's a lot of room on the upside there.
Anyway, that's why I'm running the experiment, to see how all this interacts.
|
| |
|
Edit/Delete • Quote • Complain |
newwurldmn
Registered: Apr 2011
Posts: 2617 |
03-17-12 05:40 PM
Quote from trefoil:
Good point, however I did a call spread, not an outright sell, so max risk on each done is 500-108 = 392 per. 392 * 6 = 2352 max risk overall, against max profit on the put side of 892 * 2 = 1784. Cost of the put spreads was 108 per, so assuming absolute worst case and held to expiration I'd see a loss of 2352 + 216 = 2578. Stinging but not catastrophic.
As you're probably aware, the diff between each strike on the VIX options isn't that high, and expands and contracts only slowly. So, as a practical matter, once you've done a spread the max risk on each spread would only come into play on a truly extreme and rapid move in the VIX, remembering that a rise in the implieds hits both sides of the spread, after all, firstly, and secondly the SPY put spread value would rise as well with a rise in overall implieds that a rise in the VIX is a measure of. Considering how wide I made those put spreads I'd see a nice increase in value on any increase in IV that accompanies a downmove. Overall IV on April puts according to TOS at the moment is only 15.4%, so there's a lot of room on the upside there.
Anyway, that's why I'm running the experiment, to see how all this interacts.
I like your structure.
+400ish if we rally from here.
-250 in a crash
and who knows in between. But I think it could be > +400 and unlikely to be less. Best case scenario, we grind down 3-4% and sport VIX is around 22 when we settle.
And I can't imagine a reasonable scenario where spot vix is 28 and spot is higher.
The VIX callspread helps cut the convexity risk out.
|
| |
|
Edit/Delete • Quote • Complain |
| Receive
an email whenever a new post is added to this thread by subscribing
to it. |
|
|
|
|
