"Internal Error 500 500" when Forecasting

[sarah_ec]

Hi all,

I tried to run an ex-post in sample dynamic forecast of my OLS equation, but each time I receive the error "Internal Error 500 500". I tried to google the problem but found no answer. Anyone an idea what it could be about?

Thanks to anyone who takes the time to read this!

[sarah_ec]

I just found out that Eviews actually has created the forecast, although showing me the error. I conducted the forecast again and still got the error, but the forecast was again created. I just hope that the numbers are correct...

[EViews Gareth]

What version and build date of EViews? (Help->About EViews)

[sarah_ec]

Sorry for not mentioning, it is Eviews 9.

[EViews Gareth]

What's the build date?

[sarah_ec]

May 29 2015

[EViews Gareth]

Update to the latest version.

[sm959]

Hi,

I too get an "Internal Error 500 500" when forecasting in Eviews 9 (build 7 Jan 2016).

I estimate a least squares panel with cross-section and period fixed effects. When trying to forecast, the "Internal Error 500 500" pops up, although the forecast is produced in the background. Interestingly, the error goes away when I untick the box for standard errors in the forecast specification. Equally, removing the fixed effects seems to resolve the error. Does this imply that I cannot use the normal standard errors when forecasting a fixed effects model?

As the loop in my programme breaks down due to the error, I'd be very grateful for suggestions as to how to resolve it.

Many thanks in advance,
Sam

[EViews Gareth]

Could you provide the workfile/program?

[sm959]

Thanks for your quick reply. Please find a schematic representation of the workfile attached. Running a forecast on eq01 produces the error.

[EViews Gareth]

I believe the error is caused by the lack of a constant in your equation's specification (note that EViews puts one in automatically, but that causes issues later on). We'll fix the bug, but in the mean time you can work around it by adding a constant to your specification.

[sm959]

you can work around it by adding a constant to your specification.

Great, thanks, that works!