Page 1 of 1

How to structure panel data in order to use it in a model

Posted: Tue Mar 19, 2013 4:36 am
by Raz
Hi there,

I am very new to eviews and statistical programs in general. I want to analyse a huge data set in eviews version 7.
I have watched some tutorials on youtube and began reading in the user guide, but i have come to a point where I am quite unsure what sort of keywords i need to look for in order to solve my problem.
I hope you guys can help me by giving me advice and/or pointing me in the right direction and i apologize for what might seem like "nooby" questions.

The Data
I have a data set of 300k companies between the years 2000 and 2011.
For each company I have a number of variables in the form of accounting data (Assets, Liabailites, Profit, Leverage, Listing status and so on)
I know already that this means i have either pooled or panel data.
The unique ID (for cross sections) i used when structruing the file is company ID nr and i used the year for the time series ID.
In addtion I have a parameter that provides information about what industry the company operates in (8 industry types with a numerical id between 1-8).
By taking a closer look at the data however I realized that some companies have different industry IDs in different year (which means they changed the industry sector between years).
From what i can tell this is a minority of the data set, but I am not sure if this could cause a problem later.

Data.jpg
Data.jpg (131.2 KiB) Viewed 2890 times


The Problem
1. However I would like to create a table as the one attached to this post. The table shows leverage per industry and year and in addition provides information about the mean for the the whole period.
Dokument-2013-03-19-12-14-23.jpg
Dokument-2013-03-19-12-14-23.jpg (370.29 KiB) Viewed 2890 times

Essentially i have to combine the data somehow but I am not sure if I have to do this in different steps or if there is an option to do all this in one step(?!)



2. In additon i would like to create a "financial crisis dummy" which means that the years 2000-2008 should get a value like for eks. 0 and 2009-2011 get the value 1. This is needed for my regression model i want to use later on.
Do you have some advice how to do that too :)!?

Model
(Longterm debt)/(Total Assets) = α+ β1*IML+ β2*Tang+ β3 P+ β4 FS+ β5*Growth+ β6*Ex.Inf + β6*FA+ β8*CrisisDummy + ԑ


I hope that you guys can help me with my problem and a appreciate thetime you take to look at "my challenge" :)!
Thanks in advance

Raz

Re: How to structure panel data in order to use it in a mode

Posted: Tue May 14, 2013 3:20 am
by NicV
Regarding your financial crisis dummy, a very helpful thread on dummy variables is the following: viewtopic.php?t=3840&f=4

So for your model you could create a dummy using the following code:

Code: Select all

series fincris = @recode(@year>2008, 1, 0)