Dear David:

I have one question in the parametric volatilities calculation spread sheet. You, in cell G11, use the EWMA formula and weight return and lagged variance with lambda and 1-lambda. while the variance input in this formula has been taken from cell I28 which is already the sum of weighted (with 1-lambda) variances. Do we need to weight it twice?

Secondly, there is a minor error in cell E5 that should be 2.3 %

best regards,

Peter

Hi Peter,

Note there are two versions of EWMA (cell G11 & G12), the second (G12) is the manual EWMA: each return^2 given a declining weight. In that full-blown, there will be no double-counting of weights. Okay, but the recursive EWMA (G11) is the “solved” version of the previous infinite series; or, if you like, it autoregresses on itself (current estimate is function of previous variance). So, the recursive EMWA indeed weights the lagged return^2 and the lagged variance. Because the lagged variance essentially incorporates the previous infinite series; the lagged variance is shorthand for weighting the previous squared returns. Note that I28 is the previous (yesterday’s) variance.

Re E5, sorry you may be right, however i do not see the error, E5 = SQRT(moving average variance)?

David