Can some highlight Adv/Disadv of Monte Carlo VAR?
It provides the probability of distributions of risk and return. It takes into consideration of inflows and outflows of cash flow.
Advantages: doesn’t require a normal distribution and can handle complex relationships among risks. Disadvantage: can be very time consuming and costly to implement, requires many simulations to get the model accurate, and has many inputs which must be estimated.
Adv: Ability to incorporate any return distributions and asset correlations. Does not require normal distribution although often used. (useful for derivatives dealers without historical returns and normal distribution) Disadv: Has to make thousands of assumptions about return distributions & their correlations for all inputs
More… A: - more flexible than other methodologies, because can specify distribution - can consider portfolios with investments having non-linear return distributions, e.g. putable bonds - considers all possibilities, even the most extreme situations - i.e. outliers D: - the more variables considered, the more simulations you have to run… i.e. complexity
I know it is been stated that we assume normality of returns, but in reality, returns are skewed (eg. 2008 market crash), therefore the VAR generated from Monte Carlo simulations is meaningless, therfore a disadvantage.
this is an argument against VaR generated by any means, but MC VaR is superior to delta-normal or historical computations. There are techniques such as computing tail VaR or expected shortfall that attempt to mitigate some disadvantages of VaR