EViews.com

Hello,

I have some basic knowledge of Eviews, but for my next project I'm facing some problems. I have been reading the user guides I could find, but still some unresolved questions remain.

Let me briefly explain the purpose of my project:

I have to test whether some kind of theory (the pecking order theory) is correct.
I have data for 220 companies, for 15 years and 6 variables (deltaDebt, deltaMTB, DEF, PPE, deltalogsales, profitability)
The regressionmodel(s) I want to test are (approximately) :

DeltaDebt = a + bDEF + e

DeltaDebt = a + bDEF + cPPE + ddeltaMTB + edeltalogsales + fprofitability + e

The purpose of my project is that I test whether these regressions 'are true' for every year separately, and over the 15 years together.
For every year separately I work with an excel file, with a tab for every year. And I estimate both equations (LS) for each year.

The problem is I don't really know how I can make an estimation over the 15 years together this way.
Recently I have noticed I'm actually dealing with 'balanced panel data'. So I have made a new excel file with 220 cross-sections (companies) stacked on top of each other. I believe this might be the solution to my problem, but I don't really know how to approach the analysis correctly. I have been reading about 'fixed effects', but I don't really understand what this means.

I don't know if this is very clear to you, but if you have any further questions necessary to answer this question, please ask. And if it is clear, Is there someone who knows a solution to my problem? Or who can briefly explain what my possibilities are with balanced panel data?

Thank you very very much!!

Bring it in as a panel. Then it is easy to do the 15 estimations - just change the sample for each estimation (for the first estimation, the sample would be "1990 1990", for the second it would be "1991 1991" and so on...).

I wouldn't worry about fixed effects.

Oh, thank you!

So, as I understand, if I do it as you say, I don't need my excelfile with separate tabs for every year? This way I can estimate my regression for each year separately? Will this approach give me different results than my previous approach, or is this kind of the same?
What's the added value of this panel data?

But what I still don't really understand is how I can estimate my equation for all the years together? (Thus, I want to know what (for example) my R-squared value is over the 15 years.. Or is this really the mean of all my separate values?) I want to be able to make a conclusion about 'the correctness' of my assumption in general.

I hope my questions don't sound to stupid, but I'm really new to this panel data, and I'm really trying to understand how it works and what the possibilities for my research are :)

Thank you again! :)

The added value of bringing it in as a panel is that you can choose whether to use a single year of data, or all 15 years, simply by changing the sample. Also, you don't need separate Excel sheets.

Running a regression on a single year of data in a panel workfile is the same thing as running it on a cross-section workfile, so you'll get the same results

And what does it exactly do when I use all the 15 years instead of single years?
How do I have to interpret the output in that case? (The R-squared for example?)

Thank you!

It uses all of the data rather than just part of the data.

I can't help you with interpretation.