AI need an EWMA variance/covariance matrix to calibrate a simulation. How would I go about calculating the EWMA matrix from historical stock data?Thanks!

Is this just the variance of individual stocks and the covariances between them?

Yes this is the correct formula for covariance. Of course, there is nothing wrong with what you have written, but you have used the symbol rho. Of course, this symbol is usually reserved for correlation, which as usual, is found by dividing the covariance by each of the volatilities.The formulas are in Hull by the way.If you have a long data series this formula is ok if you just want the current value. However, if you have a shorter time series (anything less than a year) or you want to look at the results of the entire time series, you need to come up with some rule of thumb for starting the volatility and covaiance estimates. These formulas start at 0, which is simply a default by omission value.