Dated panel data

[daan]

Hello,

I writing my thesis on the effect of the board of directors characteristics on CEO compensation.

My data consist of firms with each several years making it (i think) dated panel data.
When i want to test for heteroskedasticity, serie correlation and multicollinearity i search on youtube how to, but when i want to execute i don't see the same options. for example when i go to "view" -> "residual diagnostics" i don't get "serial correlation LM test" this is the same for the other tests.

The big difference between my screen and that of youtube is that they have in there equation output "methode: least squares" and i have "methode: panel least squares"

Why is this? and does it mean i don't have to test for it or that i need to use another program?

Please help!!

[daan]

Hello,

I solved this problem by uploading my data as "unformatted" instead of "dated panel data" now i can run my tests and regressions.
The question is, is this correct? can it be done this way?

thanks

[EViews Gareth]

If you don't want panel specific estimation, yes

[startz]

Hello,

I solved this problem by uploading my data as "unformatted" instead of "dated panel data" now i can run my tests and regressions.
The question is, is this correct? can it be done this way?

thanks

Probably not, at least not for the serial correlation test. When you tell EViews something is a panel it goes to all sorts of work to check observation boundaries. That doesn't happen when you say unformatted.