Rolling GARCH with constraint in the conditional variance

[ulysse]

Hey,

I am currently working on my thesis, and trying to develop a sliding
window (rolling?) GARCH(1;1) analysis on stock index data.

The idea is to perform a GARCH/TARCH analysis on a dataset of 3000
daily observations with a fixed sliding window of 300 observations. I
have the following information.

- Returns (RET)
- Volumes (VOL) - Introduced in the conditional variance with another coeficient (Gamma for example)
- COEF (To which the coeficients of the GARCH(1;1) should be
exported).

The objective of the programme should therefore be to perform a GARCH
analysis on the 300 observations of the fixed window, export the
coeficients to the COEF workfile and then repeat the task 90 times by
sliding the window by +30 observations. This should eventually produce
90 different GARCH(1;1) coefficient estimations in the COEF workfile.

Any help would be welcome since I am not used to programming in EViews
(5).

Thank you for your answers!

Ulysse

[trubador]

Similar examples.