Are simulation techniques appropriate?

One way of generating a large number of market outcomes is to use simulation
techniques, such as the Monte Carlo technique. Simulation techniques produce a
random set of price outcomes based on the market characteristics assumed. Two
points should be made here. One of the key market characteristics usually assumed
is that price changes are normally distributed. The second point to note is that
modelling extreme moves across a portfolio is not a practical proposition at present.
Standard Monte Carlo simulation models will only produce as many extreme price
moves as dictated by a normal distribution. For stress testing we are interested in
price moves of greater than three standard deviations, as well as market moves
covered by a normal distribution. Therefore, although simulation may appear to
provide a good method of producing stress test events, in practice it is unlikely to be
an efficient approach.