The 'rolling' command might work (I think its only available in
Stata 9)
but the syntax is the following:
rolling window(#) stepsize(#) saving(pathname\results.dta), <command
you want to execute, such as regress>
what this does is it takes the first 'X' observations (this is what
you specify in the <window> option) and runs <command> on them, and
saves the estimated coefficients in results.dta. It then rolls the
window forward by the number of time period specified in
<stepsize>, and then repeats the estimation again, saves again.
Recursive estimation is also possible (keep the starting point fixed
and just keep expanding the estimation window forward).
Quoting Nuno <[email protected]>:
> Hi everyone,
>
> I have a database that has the following structure:
>
> Firmid date ret mktret
> A 011990 .1 .05
> A 021990 .05 .02
> A 031990 -.05 -.1
> (...)
> A 122000 .1 .1
> B 082001 .2 .1
> B 092001 .05 .2
> (...)
> B 122006 .1 .1
> (...)
>
> I need to calculate a simple regression of ret on mktret for each
> of the
> last 24 months of data of each company, for each company, and
> save the
> results in a separate file.
> If I use the command:
> statsby _b , by(firmid) saving(c:\results.dta,replace): reg ret
> mktret
> I get the coefficients using all the available observations,
> which is not
> what I need. I've tried using
> statsby _b , by(firmid date) saving(c:\results.dta,replace): reg
> ret mktret
> in -1/-20
> But this doesn't work as the -by- statement restricts the sample
> to the
> month being analyzed, thus not including the last 20 months in
> the
> estimation.
> Any ideas of this can be done?
>
> Best,
>
> Nuno
>
> *
> * For searches and help try:
> * http://www.stata.com/support/faqs/res/findit.html
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
>
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
