<>
I still don't think you want or need an explicit loop. I presume this
is panel data from CRSP, sorted by permno and date. Generally what you
are trying to do can be done with a by-prefix (see ISP below). For
computing rolling measures in a panel, you can either use the rolling
prefix (which will create one new file) or Baum-Cox -mvsumm-:
webuse grunfeld, clear
ssc inst mvsumm, replace
mvsumm invest, gen(rollvol) stat(sd) window(3) end
l company year invest rollvol if company<4
In no case should you need to create a separate file for each permno.
Presuming you have hundreds of permnos, that is crazy. Keep the stuff
all in one file (using either rolling: or mvsumm) and pick out what
you need.
On Oct 14, 2009, at 2:33 AM, Kushan wrote:
Thanks for the response Kit.
You are right, these are permno from CRSP.
However, I am not looping over them to create another value. What I
need to do is, create another file (one per permno) that contains date
and the rolling volatility (something that I compute) for the 15 day
period.
And since I will need a separate file for every permno, I figure
looping over permno will be the best way.
What do you think?
Kit Baum | Boston College Economics & DIW Berlin | http://ideas.repec.org/e/pba1.html
An Introduction to Stata Programming
| http://www.stata-press.com/books/isp.html
An Introduction to Modern Econometrics Using Stata | http://www.stata-press.com/books/imeus.html
*
* 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/
