Correlation, co-dependency and risk aggregation [13]

Go to: Summary | Previous | Next   
Bullet points include: Two main approaches commonly used to introduce dependency between returns / outcomes. Specify a factor structure. Specify a copula. Related because: Any dependency structure can be expressed via a copula (and relevant marginal distributions), as we shall see from Sklar’s theorem. Some specifications in one manner have natural interpretations in the other. But in practice often rather different ways of introducing dependency. Factor structure usually assumes normality, but copula usually non-normal

Contents | Prev | Next | ERM Lecture Series

Desktop view | Switch to Mobile