EViews.com

Hello,

I am trying to run a FOURVARIATE GARCH BEKK MODEL, however I encountered a problem at the very last part. The error message reads ' Overflow in scalar lr = -2*(eq1.@logl + eq2.@logl + eq3.@logl + eq4.@logl - fvgarch.@logl ) ). Here the alphas and betas have an initial value of 0.1. I tried to change to alpha and beta to 0 however, then an other message reads ' positive or non-negative arguments to function expected in scalar lr_pval = 1 - @cchisq(lr,4). Would you please suggest any solution to any of these two error messages

Thank you

Error messages are pretty clear. Scalar lr either end up with a very large or a negative number. I do not think playing with the starting values would be of much help. You are trying to build a very complicated framework with monthly data. Data span is very short and y3-y4 do not even have significant univariate GARCH effects. I believe the model is ill-defined and you'll have to figure it out on your own.

Thank you for your message. Could an increase in frequency be the solution?