st: Re: rolling calendar ests

[Kit Baum <baum@bc.edu>]

You should be able to do this with mvsumm if you tsset correctly, e.g. 
generate a sequential date variable that counts the number of trading 
days, and then use that variable (and your firm indicator) in tsset. 
Stata does not really understand business-daily data, so you have to 
create a fake time variable that counts trading days.

Kit Baum, Boston College Economics
http://ideas.repec.org/e/pba1.html

On Mar 23, 2005, at 2:33 AM, Yvonne wrote:


> However, if I wish to do a rolling estimate giving the standard 
> deviation of
> the last actual rather than calendar year (e.g. if today is 22 March I 
> want
> the standard deviation from 22 Mar 04 to 21 Mar 05, not from 1 Jan 04 
> to 31
> Dec 05), is there an easy way to do that? If I know there are 252 
> trading
> days per year, can I just ask it to calculate the std dev of the last 
> 252
> options? I'd like to do something like
>
> egen sigma = sd(ret[_n-252]:ret[_n-1]) if firm[_n-252] == firm[_n]
>
> but it doesn't seem to allow anything like that.
*
*   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/