IMPLIED CORRELATION - HOW?

derivababy · November 6th, 2003, 5:38 pm

Come on,anyone else to explain that thing....

FDAXHunter · November 6th, 2003, 7:37 pm

Think about it this way: If you are reseting every period, then you want the components to disperse to the maximum amount, i.e. if 1 is up, the other is down, and vice versa. That way, you are guaranteed that there will be a performer for that period (you are looking for upside only participation).Makes sense?

Sthrnplayalistic · November 12th, 2003, 3:14 pm

True, but that doesn’t yield that if performance is measured from inception a himalaya is long correlation. At first the himalaya is short correlation, once started I don’t know. Maybe it depends on where the spots are relatively to each other and we could go from short to long correlation at some point (as a buyer).

Sthrnplayalistic · November 12th, 2003, 3:21 pm

oops, forgot that best performer is removed from the basket... in that case, yes, long corr...

FDAXHunter · November 12th, 2003, 3:26 pm

This is true. However, at the outset, it is long correlation, especially if you introduce a floor on the sum rather than individual sums. Obviously as stuff moves around it can change. Generally, the more local features (reset, floors) you introduce, the more sensitive to correlation (with respect to the different underlyings and vols, etc) the thing becomes. As you rightly pointed out, it can move around quite drastically then.

Rouletabille2 · November 12th, 2003, 3:45 pm

FDAXHunter · November 12th, 2003, 3:48 pm

Consultant... you mean like you?Cheers.

Rouletabille2 · November 12th, 2003, 4:43 pm

Sofiane · November 12th, 2003, 5:37 pm

there is a small risk playing with implied correlations is obtaining correlations values >1 or -1< !

FDAXHunter · November 12th, 2003, 7:41 pm

Sofiane, there is no risk. If you obtain an implied corr. of -1< or >1. That's an arb.Regards.

FDAXHunter · November 13th, 2003, 12:51 pm

Rouletabille: You should sell your posts for £10 each, you'd be rich.34,450 GBP is rich to you?! Hm, to you, it actually might be, who knows.