There has been a lot of literature recently regarding dynamic correlation (Engle 2002, etc. ) Is there any way to get a closed form solution to a hedge on it? I am guessing that using DCC would imply probabilistic measures that would preclude the construction of an equivalent martingale measure, but I am holding out hope. I tried looking at some things from Fixed Income like Jimmy Hilliard and others on basis risk, but I didn't see anything that captured dynamic correlation. Obviously, if there is anything on dynamic beta hedging I would like to hear about that as well but I just don't see how to hedge covariance in a closed form.Any suggestions?

I think Engle's DCC is an econometric model, it depends on the dynamic feedbacks of asset innovations, which is hard to reconcile in the risk neutral world. However, this web link has details on how to dynamically hedge options:http://www.thetaris.com/wiki/Hedging_in_ThetaML

Thanks for the link.I have looked at things from Carr-Madan, Bossu, etc. and everyone seems to use average correlation or covariance. I am trying to do this for the commodity space so my hope is that there may be some way to construct it using spread options or some other rainbow variant. I suppose that worrying about DCC vs. some stochastic process is all but irrelevant at this point.

hello gjbiren. are the Carr-Madan and Bossu's papers available online?

Carr and Madan is a Risk article: http://www.math.nyu.edu/research/carrp/ ... r.pdfBossu has quite a few: http://math.uchicago.edu/~sbossu/Correl ... .pdfDerman also weighed in with outperformance options: http://www.ederman.com/new/docs/gs-outp ... tions.pdfI am trying to find a link for Kirk and Aron 1995 "Correlation in the Energy Markets" so if anyone has it, I would appreciate it.

thank you for the references. i admit i'm not familiar with these Carr, Madan and Bossu's writings.

Are you trying to hedge the cross-gamma in a model with dynamic (= stochastic) correlation?In this case I don't think you can do that without dynamically trading another option, which would be too expensive to executeWhat is your ultimate purpose?SB

Phaedo-Sorry for the long response time. What I am actually working on are variance swaps. If I recall correctly from Carr et al., following a Taylor expansion, the third term, which should be negative and approximate the variance risk premium, is a function of the correlation/covariance. I am ultimately working on a strategy using a spread between variance swaps and the best approximation I can think of is the correlation, or acrtually a function of it.GB

QuoteOriginally posted by: gjbirenThanks for the link.I have looked at things from Carr-Madan, Bossu, etc. and everyone seems to use average correlation or covariance. I am trying to do this for the commodity space so my hope is that there may be some way to construct it using spread options or some other rainbow variant. I suppose that worrying about DCC vs. some stochastic process is all but irrelevant at this point.Hello,You can definitely do it with exotics options like spread options but you won't find any liquidity at all on commodities for such products so it'll cost you a lot, because exotic market-makers cannot hedge themselves also, just maybe by controlling the flow of correlation relative to client needs but whatever you won't be able to re-balance your hedge.On commodities, if it's not vanilla then, be ready to pay honey.