Frank Copula vs. Gauss Copula

REden · July 24th, 2006, 9:52 am

I am working on a comparison between different copulas, but don't see an major difference between the Gauss and the Frank Copula. Tail Dependence of both copulas equal zero and also the bivariate scatter plots look not very different to me. Where is the difference between the generation of the two dependence structures?Would be happy, if you can give me a hint.REden

jhlm · July 26th, 2006, 10:04 am

To see the difference between the frank copula and the gaussian try to enhance the parameter of the frank copula. To get more intuition remember to look at the distribution function - not the copula in isolation (which in itself is a distribution function). I have attach a picture showing 4 2-dimensional distributions function with different copulas. To me the dependence induced by the frank copula looks more like a butterfly (for large parameter) where the gaussian looks like a cigar again for large parameter Hope it helps/Jonas

Kommander · March 28th, 2008, 9:23 am

Hi Paul you could use a goodness-of-fit test, if you have existing data to fit. See for examplehttp://www.uni-konstanz.de/micfinma/conference ... Bakken.pdf