Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: rollreg

From   Nick Cox <>
Subject   Re: st: rollreg
Date   Tue, 1 May 2012 18:25:02 +0100

I am not sure how much faith to put in standard deviations based on 3
values, but you don't need the looping of -rolling- here

local y roe
gen sd = sqrt(((`y'^2 + L.`y'^2 + F.`y'^2) / 3) - ((`y' + L.`y' + F.`y')/3)^2)

subject to whether you prefer a different divisor.


On Tue, May 1, 2012 at 5:40 PM, "Fabian Schönenberger" <> wrote:
> Dear Statalist
> In order to conduct moving standard deviations for the return on equity for my panel data I used the following command:
> rolling sd=r(sd), window(3) keep(cusip) clear: sum roe
> Data sample is quite large, ca. 250'000 yearly-observations, ca. 8'000 companies. Unfortunately, the command takes several hours to produce the output. Is there any other possibility to conduct moving standard deviations for unbalanced panel data?

*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index