Panel data export

[jaromir]

Hello,
I have been making my first steps with EViews programming and data exporting.

For panel models with fixed cross-section effects, you can display the values of the cross-section effects in model view options. My question is: is there a way to export these data to a table (or - I believe I've seen it - to a series) using programming?

Regards and thank you

[EViews Gareth]

You can use the @effects data member of your equation to save them into a vector:

Code: Select all

vector cx = eq1.@effects


[jaromir]

Thank you.

Does this feature works for equation objects (as well as pools)? I keep getting a "Syntax error" when I try to use it...

Regards

[EViews Gareth]

Is your copy of EViews up-to-date?

[j.galimberti]

Alternatively, you can do it without the new function:
1 - Fit the equation to a temporary series:
eq01.fit(u,f=na) tmp_series
2 – Lets say your equation is specified as: y=c(1)+c(2)*x1+c(3)*x2. Then you can generate a series with the fixed effects with:
genr fix_effects=tmp_series-(eq01.@coefs(1)+eq01.@coefs(2)*x1+eq01.@coefs(3)*x2)

The generated series will have the fixed effects (I didn’t tested for random effects) that you have specified in your equation. However, if you specified cross-section and period effects, the series will represent the sum of both.

[Dusan]

Hi,
I m running a loop of multiple regressions and I only need the R2( R squared) of them in a matrix, vector, excel or whatever!
I would be really greatful if u could help me on this.
(Till now the only thing that I succeeded was to have a matrix(or vector) for each R2.So, alternative if u could help me on how I can transform all these vectors(matrixes) in one it would also be perfect.)

[j.galimberti]

If I correctly understand what you want I would do something like this:

Let's say you have a table with each row representing an equation that you would like to estimate, and each column representing a variable. The first, obviously, is the dependent variable. Name this table as "specifications"

!n=1
while @isempty(specifications(!n,1))=0
%comand=""
!n2=1
while @isempty(specifications(!n,!n2))=0
%comand=%comand+specifications(!n,!n2)+" "
!n2=!n2+1
wend
equation eq1.ls {%comand} ''' Here is the estimation
vector(!n) r2s
r2s(!n)=eq1.@r2
!n=!n+1
wend

This program will create a vector called r2s with the r-squared of each specification contained in the table "specifications".

I hope it helps.