Hi,
I must admit I'm a bit struggling with the copula study..
Could anyone shed some light for me / correct my understanding.
- The copula is useful as it allows to build a single multivariate distribution from several single variable distributions.
- To do this, we map the values of the single distributions to values of the multivariate distribution using the probability mass of the respective distributions (this is fairly well illustrated in Figure 10-2 of the GARP reading) I assume..
- By I do not know which way, this multivariate distribution also assign a correlation to (?)
- Then what?
I am struggling to see what's the interest for this?
They give an example which is: What is the probability for 2 companies to default:
Correlation = 0.4
PD A= 6.51%
PD B= 23.83%
They give as an answer 3.44% but trying to calculate I got:
Default of both A and B = 6.51% x 23.83% x 0.4 x Sqrt [ 6.51% x (1-6.51%) x 23.83% x (1-23.83%) ] = 5.75%
I would be very grateful if someone could clarify this for me!
Thanks.
I must admit I'm a bit struggling with the copula study..
Could anyone shed some light for me / correct my understanding.
- The copula is useful as it allows to build a single multivariate distribution from several single variable distributions.
- To do this, we map the values of the single distributions to values of the multivariate distribution using the probability mass of the respective distributions (this is fairly well illustrated in Figure 10-2 of the GARP reading) I assume..
- By I do not know which way, this multivariate distribution also assign a correlation to (?)
- Then what?
I am struggling to see what's the interest for this?
They give an example which is: What is the probability for 2 companies to default:
Correlation = 0.4
PD A= 6.51%
PD B= 23.83%
They give as an answer 3.44% but trying to calculate I got:
Default of both A and B = 6.51% x 23.83% x 0.4 x Sqrt [ 6.51% x (1-6.51%) x 23.83% x (1-23.83%) ] = 5.75%
I would be very grateful if someone could clarify this for me!
Thanks.
