Estimating an ARCH model

For technical questions regarding estimation of single equations, systems, VARs, Factor analysis and State Space Models in EViews. General econometric questions and advice should go in the Econometric Discussions forum.

Moderators: EViews Gareth, EViews Moderator

mvm20
Posts: 2
Joined: Mon Aug 19, 2013 7:31 pm

Estimating an ARCH model

Postby mvm20 » Mon Aug 19, 2013 7:50 pm

Hey all,

I am having trouble running an ARCH model, I would like to model a fisher relation except I want to add an extra term, the conditional variance of inflation.
If my model is then interest = c + inflation + var(inflation) how do I model var(inflation) before I put it in the equation?
Thank you in advance, I am having trouble and any tips or help would be much appreciated.

Cheers.

trubador
Did you use forum search?
Posts: 1518
Joined: Thu Nov 20, 2008 12:04 pm

Re: Estimating an ARCH model

Postby trubador » Thu Aug 22, 2013 2:39 am

This is an example of garch-in-mean model. Open the "Equation Estimation" window and select ARCH as the estimation method. You'll see an "ARCH-M" option to the right of the "Mean Equation" dialog box. Simply select the "Variance" option from the drop-down list. I suggest you to try other specifications as well and see which one fits the best.

mvm20
Posts: 2
Joined: Mon Aug 19, 2013 7:31 pm

Re: Estimating an ARCH model

Postby mvm20 » Wed Aug 28, 2013 4:18 pm

Thank you trubador,

When I run this model and extract the arch terms by using "proc->make GARCH variance series", what is the difference between selecting variance in "arch-m" and not selecting anything at all?
Sorry I am rather new to econometrics.
Also in the mean equation window should I put the interest rate and its lags, or just interest = c + inflation? Because the latter says near singular matrix. And one more thing, does anything need to be put in the variance regressors box?

Thank you very much for your help.


Return to “Estimation”

Who is online

Users browsing this forum: No registered users and 14 guests