GARCH(1,1) with constraints

[DiederikvanGemert]

Hello everybody,

I am modelling volatility through different GARCH models, in order to find the best fitting volatility for a VAR estimation. I am quite new to EViews and have been looking around this forum for an answer to my question, unfortunately I haven't found an answer to my question.

When doing a GARCH(1,1) on NASDAQ returns I get a result where (A+B)>1. I read on this forum that I should impose restrictions on alpha en beta by adding exp(alpha) and exp(beta) but then I get a near singular matrix (see image). My equation is as follows: Y c AR(1). How can I get my combined alpha and beta to be lower than 1? Same holds when I am performing an EGARCH model.

Any help is appreciated.

[trubador]

1) A+B>1 is a condition to check the stability of your GARCH model ex-post.
2) GARCH model requires the residuals of the mean equation to be stationary.
3) You will soon discover that the return formula of nasdaq_100 is incorrect.