If somebody would have any time just to help a bit for this problem which could be easy to advanced user of eviews but does not open up to me even after reading the manual, books and videos.
Im estimating stock returns and extreme stocks returns effect to gold and as far as I can understand this is done with Garch (1,1) estimation.
The model I am using is:
I have used this code to derive the stock returns to the 1%, 5% and 10% biggest declines:
series d01 = stocks<@quantile(stocks, .01)
series d05 = stocks<@quantile(stocks, .05)
series d10 = stocks<@quantile(stocks, .10)
Then from a post I already saw there was this equation posted:
gold = c(1) + c(2)*stocks + c(3)*d10*stocks + c(4)*d05*stocks + c(5)*d01*stocks
and from here I would need help. Should this code be inserted straight to the mean equation, which I have done and got some mixed results. Or should some of this be inserted maybe to the variance regressors? Or should I use this @expand function somehow to the mean equation?
The biggest doubt I have is that this gives me very unsignificant coefficient for the whole period effect of stocks to gold. I use 15 years of daily data.
Also as a quick question, is this basically the same model(exclude the bond coefficients) they also use GARCH for the heteroscedasticity:
If somebody would have any time to look at this and give some help, I would appreciate it very much! Thanks in advance!