425636_en-US

liliconcom

Panel Least Square method for Panel Data - Need Help

Hi Team,

One of the professors sent me the data and output of Eviews. He has a unbalanced Panel data  and applied the Panel Least Squares using E Views where he defined Company as cross section variable and Year as a Time component.

He was trying to replicate the E Views results using JMP. (mixed Model)

I don't see anything related to Fixed effects or Random effects here and it just says Panel Least Squares. I initially assumed it to as pooled regression. 

Appreciate your help and advice.

(Attached JMP Data set and E Views Result)

Dependent variable = ROA

Independent Variables are = CP, CR, Size, LEV, and EG

Thanks

Murali

Re: Panel Least Square method for Panel Data - Need Help

I think it's a one-way fixed effect model with the company being the fixed effect. JMP can fit the model to match EViews'  with the exception of the intercept (see the screen shot below--the right panel is the JMP report). 

In EViews, C,  the  constant term,  is parameterized so that the fixed effects estimates are relative to the constant term and add up to zero, while in JMP's default parameterization the intercept term represents the estimated baseline-line "average company" of Y.

Hope this makes sense.

Re: Panel Least Square method for Panel Data - Need Help

Thanks a lot Jiancao.