simple OLS regression- F-statistic is not reported

gobbble · Postby **gobbble** » Wed Aug 01, 2012 1:00 pm

Hi,

I have the problem that at the very basic OLS regression of a return variable against its first lag no F-statistic is reported! What can be the problem and how can I get the f statistic and probability. The regression looks like "r(-1) c r"

edit: I am using eviews 7

Thanks

startz · Postby **startz** » Wed Aug 01, 2012 2:23 pm

That's a regression of lagged r on r, not the other way around.

Past that you may want to post your output.

gobbble · Postby **gobbble** » Thu Aug 02, 2012 1:01 am

here it comes. Regards

startz · Postby **startz** » Thu Aug 02, 2012 6:09 am

Something's really, really wrong as you also have a negative R-sqr. I think you've hit a bug. The EViews folks are going to want to know the version and build date of your copy of EViews (Help/About).

gobbble · Postby **gobbble** » Thu Aug 02, 2012 9:15 am

thanks. I have eviews 7 standard edition Jan 7 2010 build.

By the way, when I regress the other way around, everything turns out fine...

startz · Postby **startz** » Thu Aug 02, 2012 9:17 am

You should probably push the update button, since your copy is more than two years out of date. If convenient, it would also be helpful if you posted your EViews workfile.

EViews Gareth · Postby **EViews Gareth** » Thu Aug 02, 2012 9:19 am

You should update to the latest version, by running the upgrade installer.

http://eviews.com/download/download.html

EViews Glenn · Postby **EViews Glenn** » Thu Aug 02, 2012 9:47 am

I was waiting to see the output before I nosed in on this one, but so far as I can see, the F-statistic is always reported in this setting in the shipping version of EViews.

gobbble · Postby **gobbble** » Thu Aug 02, 2012 12:20 pm

unfortunately I cannot uldate it, asI use the license of my university. I have admin rights to update it...

EViews Gareth · Postby **EViews Gareth** » Thu Aug 02, 2012 12:21 pm

Post your workfile, so we can check it has been fixed in the latest version for you.

gobbble · Postby **gobbble** » Fri Aug 03, 2012 3:29 am

here is the workfile

startz · Postby **startz** » Fri Aug 03, 2012 7:07 am

I can verify that the bug has not been fixed.
So that you can get on about your work, I have done the following.

Code: Select all

series rlag=r(-1)
ls rlag c r

I've posted the results for you, but I'm 99% certain that you want

Code: Select all

ls r c rlag

instead.

: Capture.PNG (23.02 KiB) Viewed 14897 times

EViews Glenn · Postby **EViews Glenn** » Fri Aug 03, 2012 9:55 am

Thanks for posting the workfile gobble. That helps a great deal.

That's a good suggestion startz and the one that I've used to verify the issue.

I'll have a fix posted shortly.

[edit: it's up]

EViews Glenn · Postby **EViews Glenn** » Fri Aug 03, 2012 12:02 pm

A fix has been put into the main code base. Please look for the update shortly. I'll try to get them to get it out today.

gobbble · Postby **gobbble** » Sat Aug 04, 2012 1:41 am

thanks for your help!

EViews.com