• C++ Programming for Financial Engineering
    Highly recommended by thousands of MFE students. Covers essential C++ topics with applications to financial engineering. Learn more Join!
    Python for Finance with Intro to Data Science
    Gain practical understanding of Python to read, understand, and write professional Python code for your first day on the job. Learn more Join!
    An Intuition-Based Options Primer for FE
    Ideal for entry level positions interviews and graduate studies, specializing in options trading arbitrage and options valuation models. Learn more Join!

Marginal VaR of Swap position

Joined
5/6/06
Messages
384
Points
28
Is it possible VaR of a single name swap can increase total VaR in a portfolio. I don't quite sure where the risk comes from as to a SWAP...:sos:
 
Hi,

It is certainly possible. Single name equity swap? The marginal VaR is usually calculated with the covariance matrix E -- the portfolio variance is w'Ew, and then you scale up the square root of that by some coefficient to get portfolio VaR (95%, 99%, 1 day, 10 day, etc).
The derivative of this portfolio variance is a vector 2Ew, which should be the vector of marginal variances from each position. From there you work back to marginal VaRs.

For single name credit default swaps or asset swaps, the covariance matrix method is not used as much. Monte Carlo is another way to get VaR, but it is more difficult to get marginal contributions -- you can run 2 portfolios on the same random paths, one portfolio with, and one without, the swap, and the VaR difference might be considered "incremental VaR" of the swap. But that is pretty costly for such a number.
 
I don't quite sure where the risk comes from as to a SWAP...:sos:
if your counterparty defaults, you are stuck holding a leg of a swap and not getting anything.
 
Hi,

It is certainly possible. Single name equity swap? The marginal VaR is usually calculated with the covariance matrix E -- the portfolio variance is w'Ew, and then you scale up the square root of that by some coefficient to get portfolio VaR (95%, 99%, 1 day, 10 day, etc).
The derivative of this portfolio variance is a vector 2Ew, which should be the vector of marginal variances from each position. From there you work back to marginal VaRs.

For single name credit default swaps or asset swaps, the covariance matrix method is not used as much. Monte Carlo is another way to get VaR, but it is more difficult to get marginal contributions -- you can run 2 portfolios on the same random paths, one portfolio with, and one without, the swap, and the VaR difference might be considered "incremental VaR" of the swap. But that is pretty costly for such a number.

Thanks Pat.
Actually, this is TRS that deals with JP Morgan. I am using historical VaR instead of Delta Normal or MC now. So you are saying the actually risk is the same as holding underlying ? :wall
 
It depends on the payment formula. For a 1 day 99% VaR, take the formula for the floating payout and figure how much you would have made/lost on the swap every day over the past X years. Then take the 99% worst day.
 
Back
Top