Simulating the interdependence of risks - ACTED notes

Discussion in 'SP9' started by Edwin, Sep 19, 2014.

  1. Edwin

    Edwin Member

    Two questions;-

    1) In the Acted course notes one of the advantages of MC Simulation is stated to be that;-


    What I don't understand is;- does Cholesky Decomposition use the correlation matrix between the simulated distributions or does one have to take the pain and calculate correlation from past data?

    Given the fact that the correlation matrix from observed values is more accurate, then is the point above that says "it is possible to simulate the interdependence of risks" really an advantage if we don't use the simulated interdependencies?

    Help will be appreciated.
     
  2. Oxymoron

    Oxymoron Ton up Member

    I'm not writing ST9 but I work with the above techniques. So that gives me dubious authority to answer you.

    Cholskey decomposition is used to make the correlation matrix positive and semi-definite. You will need to feed in the calculated correlation into the Cholskey decomposition formula. Calculation of correlation can be based on historical data or any other method you find suitable.

    Once the matrix is positive and semi-definite, we use Copula to simulate correlated random variables and hence the correlated series. This captures interdependence of risks in the simulated series.
     

Share This Page