Different R squared value

For technical questions regarding estimation of single equations, systems, VARs, Factor analysis and State Space Models in EViews. General econometric questions and advice should go in the Econometric Discussions forum.

Moderators: EViews Gareth, EViews Moderator

Khawla
Posts: 10
Joined: Sat Jan 14, 2023 1:36 pm

Different R squared value

Postby Khawla » Sat Jan 21, 2023 10:26 am

Hi,
I have applied the command ls y c x ar(1) using Eviews, and I applied the cochrane orcutt function using R package and excel.
I got the same results for all the statistics in all these applications, accept the value of R squared, where it is different in Eviews. I have attached the results.
Could you please help me to know why, especially, the difference is large.
Attachments
Results_Eviews_and_R_package.xlsx
(9.82 KiB) Downloaded 116 times

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 13294
Joined: Tue Sep 16, 2008 5:38 pm

Re: Different R squared value

Postby EViews Gareth » Sat Jan 21, 2023 6:14 pm

I don't see the AR results in your R output.
Follow us on Twitter @IHSEViews

startz
Non-normality and collinearity are NOT problems!
Posts: 3775
Joined: Wed Sep 17, 2008 2:25 pm

Re: Different R squared value

Postby startz » Sun Jan 22, 2023 8:33 am

What command did you use in R? I don't think

Code: Select all

lm()
does Cochrane-Orcutt, does it?

(By the way, I'd be willing to wager that the R-squared shown in the R output is from the quasi-differenced variables.)

Khawla
Posts: 10
Joined: Sat Jan 14, 2023 1:36 pm

Re: Different R squared value

Postby Khawla » Sun Jan 22, 2023 9:50 am

EViews Gareth wrote:I don't see the AR results in your R output.


Hi, I appreciate your quick response, thank you.

I used the function cochrane.orcutt (lm function) from the packages("orcutt"). This R function directly presents the new results of the Regression analysis using the transformed data (transformed using AR(1)) which what I had attached.

Regards,
Khawla

Khawla
Posts: 10
Joined: Sat Jan 14, 2023 1:36 pm

Re: Different R squared value

Postby Khawla » Sun Jan 22, 2023 9:58 am

startz wrote:What command did you use in R? I don't think

Code: Select all

lm()
does Cochrane-Orcutt, does it?

(By the way, I'd be willing to wager that the R-squared shown in the R output is from the quasi-differenced variables.)

Hi,
I used the function cochrane.orcutt (lm function) from the packages("orcutt"). I first applied (lm) function and then cochrane.orcutt (lm function).

startz
Non-normality and collinearity are NOT problems!
Posts: 3775
Joined: Wed Sep 17, 2008 2:25 pm

Re: Different R squared value

Postby startz » Sun Jan 22, 2023 12:26 pm

The Cochrane-Orcutt package is returning the R-squared on the quasi-differenced model. This is a bug. EViews has it right.

(Orcutt would not have been pleased.)

Khawla
Posts: 10
Joined: Sat Jan 14, 2023 1:36 pm

Re: Different R squared value

Postby Khawla » Sun Jan 22, 2023 3:06 pm

[quote="startz"]The Cochrane-Orcutt package is returning the R-squared on the quasi-differenced model. This is a bug. EViews has it right.

(Orcutt would not have been pleased.)[/]

Thanks for your response, it was very helpful.

Best wishes,
Khawla

Khawla
Posts: 10
Joined: Sat Jan 14, 2023 1:36 pm

Re: Different R squared value

Postby Khawla » Tue Jan 24, 2023 11:59 am

startz wrote:The Cochrane-Orcutt package is returning the R-squared on the quasi-differenced model. This is a bug. EViews has it right.

(Orcutt would not have been pleased.)

Hi and sorry to bother you again.
I just wonder, with regard to the use of Cochrane-Orcutt procedure and as much as I understand that if the all the dependent and the independent variables are transformed using rho, then thevalues of the predicted, residuals, and all the statistics including R squared value will be based on these inputs and be of the same form, but you said that quasi-differenced model is not be used in Eviews. My question is how the results of predicted, residuals, R squared values are obtained in Eviews if it does not depend on quasi-differenced.

regards

startz
Non-normality and collinearity are NOT problems!
Posts: 3775
Joined: Wed Sep 17, 2008 2:25 pm

Re: Different R squared value

Postby startz » Tue Jan 24, 2023 12:08 pm

EViews doesn't exactly do Cochrane-Orcutt. By default estimation is by maximum-likelihood. The difference is likely to be very small.

Khawla
Posts: 10
Joined: Sat Jan 14, 2023 1:36 pm

Re: Different R squared value

Postby Khawla » Tue Jan 24, 2023 1:02 pm

startz wrote:EViews doesn't exactly do Cochrane-Orcutt. By default estimation is by maximum-likelihood. The difference is likely to be very small.

I really appreciate your help. I would be grateful if you recommend a source.

Best Wishes

startz
Non-normality and collinearity are NOT problems!
Posts: 3775
Joined: Wed Sep 17, 2008 2:25 pm

Re: Different R squared value

Postby startz » Tue Jan 24, 2023 1:06 pm

The EViews help system describes the alternatives available in EViews quite extensively.

Khawla
Posts: 10
Joined: Sat Jan 14, 2023 1:36 pm

Re: Different R squared value

Postby Khawla » Wed Jan 25, 2023 9:42 am

startz wrote:The EViews help system describes the alternatives available in EViews quite extensively.

Thank you very much.

Regards


Return to “Estimation”

Who is online

Users browsing this forum: No registered users and 17 guests