We discuss nested simulation methods for computing loss distributions and Value‐at‐Risk (VaR) for a portfolio containing CDO tranches as well as single‐name CDS and bonds. When these instruments are modeled
via
correlated stochastic default‐intensity models, estimation of VaR requires a nested simulation procedure: in the outer step of the simulation one draws realizations of all default intensities up to the horizon, and in the inner step one uses simulation to reprice each instrument in the portfolio at the horizon conditional on the realized intensities. We analyze how a fixed computational budget may be allocated to the inner and the outer steps to minimize the mean square error (MSE) of the resultant estimator. We also review how importance sampling may be used to increase the efficiency of the inner step.