Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: granger test on panel data


From   "Richard Harvey" <richardharvey2008@googlemail.com>
To   statalist@hsphsun2.harvard.edu
Subject   st: granger test on panel data
Date   Sat, 13 Sep 2008 23:38:00 +0100

hello statlisters,

I have a panel set up like this

firm1 month varX varY varZ....
.....
.....
firm2 month varX  varY varZ....

the panel is unbalanced.  varX is a continous variable while varY has
lots of zeros, but  when present they are quite large (000s).The other
vars are the variables like size etc.
The months are all relative to some other date, which is particular to
each company . So they go from say -36 to +36.

I am basically interested in exploring the relationship between varX
and varY in the -36 to 0 month period and 0 to +36 month period.
particularly which leads or follows which. I am thinking of granger causality.

But can one do that on panel data in stata?

In this context would the large number of zeros be an issue?

There is also the issue of the skewness of varY which I hope to get
around this ( as many papers have done in the past) by log
transforming (1+ varY). Is this a reasonable thing to do?


any  help or pointers much appreciated.

thanks
rich
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index