Correlation, co-dependency and risk aggregation [31]

Go to: Summary | Previous | Next   
Bullet points include: Gaussian copula. Here, cdf of a standard multivariate normal distribution function is Nm(mu,rho) where mu = 0 and rho (i.e. rho i,j) is a correlation matrix (so e.g. rho i,i = 1 and -1 <= rhoi,j <= 1). Special case where rho i,j = 0 if i <> j is Independence or Product Copula: If rho i,j = rho for all i <> j then Gaussian copula is exchangeable. c.f. CDO pricing etc.

Contents | Prev | Next | ERM Lecture Series

Desktop view | Switch to Mobile