Volatility of the diference between prices of two assets

ashu · March 5th, 2003, 12:54 pm

Hi , I am trying to estimate the volatilty for the spread of two asset prices. ......The underlying prices are of two commodities and the spread varies from -ve to +ve values.................when i try to calculate the volatility of the returns (to be used in the option valuation) , i am unable to do so as some of the values of spreads are -ve....................how to calculate the volatility in such a case?

FDAXHunter · March 5th, 2003, 1:04 pm

Why can't you estimate the volatility of the returns? The Log() function blowing up on you? Simply use normal returns (Today-Yesterday)/Yesterday. Commodity spreads aren't usually lognormally distributed anyway.

ashu · March 5th, 2003, 1:14 pm

thnks.............the actual historical distribution is not normal.....................it is +vely skewed in some spreads and bimodal or multimodal in other cases.............in the case of simple returns, the value of return blows up when the absolute value becomes low.......this happens typically when the spread goes from +ve to -ve.....................giving values of 500% and 600% ......which is not realistic..............Anyway, is ther any otherway to model a spread?

FDAXHunter · March 5th, 2003, 1:22 pm

How about using the two vols of the assets and their correlation. That's simple enough.

ashu · March 5th, 2003, 1:43 pm

how to use it if the distribution of the underlying assets are bimodal or multimodal?.............

pareshgokhale · March 8th, 2003, 6:24 am

How about using a Chebyshev's distribution? Either for each of the assets or on the spread between the two.

newton · March 8th, 2003, 1:57 pm

ashu,You may want to look at modeling the spread as cointegration between the asset prices (when the spread is mean reverting).