Serving the Quantitative Finance Community

 
User avatar
xpatagon
Topic Author
Posts: 0
Joined: June 1st, 2011, 1:31 pm

EWMA and Monte Carlo

April 9th, 2014, 6:52 pm

As I understand it the "typical" EWMA decay factor of 0.94 was calculated by JP Morgan in riskMetrics assuming a drift factor of zero, permitting them to relate the squared daily returns and variance and hence minimize the squared deviation of the difference over a range of asset classes.Suppose that I have a Monte Carlo algorithm that models a risk neutral random walk in order to price a derivative. In this case the drift term is based on the risk free rate, including in the limit as delta time approaches zero. On the contrary, my volatility explicitly assumes that the drift approaches zero as the time difference approaches zero.Is it still justifiable to use 0.94 as a decay factor in these circumstances, or should a new factor be calculated redoing the minimization assuming a drift of the risk free rate in order to identify the optimum decay rate for the purposes of the simulation?Thanks