Testing if my errors are student t distributed

[strypste]

Hi

I estimated my model with Maximum Likelihood assuming a student t distribution. Now I would like to test if my errors are indeed t distributed. How can I test this in Eviews?

Steven

[EViews Gareth]

Open up the residual series, click on View->Descriptive Statistics->Empirical distribution tests.

[strypste]

Thanks for your reply.. However, the reason for my question is that the student t distribution is not in that list...

[EViews Gareth]

Ah, yes. That's what I get for answering before breakfast. You can do a Quantile-Quantile plot against the t-distribution to get an eyeball of how close the distribution is.

[strypste]

Thanks!

I managed to do it manually. However, if I want to include it in a program, then my command always plots the data against the normal distribution instead of the t distribution... This is my code:

growth.qqplot theory(dist=tdist)

where growth is the series with the data.

What am I doing wrong here?

[EViews Glenn]

Looks like a bug. Let me check into it.

[edit: I've put a fix into the main code base. It should show up in the next update.]