Pages

Sunday, February 17, 2013

Copula theory

A copula from two, independent standard Gaussian variables; from 0 standard deviations for each variable from the center:

A copula from two, independent standard Gaussian variables; from 2 standard deviations for each variable from the center:

Note that Guassian copulas suffer from a lack of tail dependence.  Also the difference between the copulas shows a pattern that is more minimal, the further from the center one begins their analysis.  For example, the copula from the shifted center shows a much weaker marginal tail distribution.  Notice the surface plot below, which differences the second distribution chart above, from the first distribution chart above:

2 comments:

  1. Replies
    1. Thanks much acronym. Also, given recent market volatility, you may enjoy the next article on fundamental "Implied volatility regimes". Or you can go here: http://statisticalideas.blogspot.com/2013/02/implied-volatility-distribution.html.

      Delete