The structure Monte Carlo (SMC) approach to estimating VAR simulates portfolio or asset returns using a stochastic process:  s_t+1,i = s_te^{µ + σ × z}.  

Z in the formula is a random draw from a normal distribution.  An advantage to the SMC approach is that multiple risk factors can be modeled by assuming an underlying distribution and incorporating correlations among assets. A disadvantage is that inaccurate future volatility forecasts may occur.These inaccurate forecasts cannot be improved by running a greater number of simulations. Also, if the covariance matrix used to model the returns is drawn during normal times, then the SMC approach will not accurately predict a scenario involving a correlation breakdown (sudden increase in volatility coupled with a sudden increase in correlations). Increasing the number of simulations does not help solve the correlation breakdown problem either.