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]

st: clarity on varsoc command

From   "Michael Ndlovu" <>
To   <>
Subject   st: clarity on varsoc command
Date   Wed, 15 Feb 2012 11:03:36 +0200


I am trying to create a time-series model with the optimal lags for each variable. I am using Stata version 12.

My understanding of the "varsoc" command is that it gives you optimal lag choice for a set of variables. So if I have type in the command "varsoc y x z" , I will get the optimal lag choices for a model that looks like: y = cons + b1*ylag + b2* xlag + b3*zlag.

My problem with this though is the assumption of the same lag period for all these variables. How do I get optimal lag choice for each variable such that the optimal amount of lags to y is not necessarily the same as x or z. 

Is it correct for me to isolate each variable and make the other variables exogenous

i.e. varsoc y x, exog(z)
      varsoc y z, exog(x)

I just need clarity on whether I am going about this the right way or whether there is another way of finding what I want.  


--------------------- *** Disclaimer *** ----------
This e-mail and its contents are subject to the SA Reserve Bank's
Disclaimer and Confidentiality Clause, which can be viewed at:
Should you be unable to access the link provided, please send a
blank e-mail to
--------------------------- *** Disclaimer *** -----

*   For searches and help try:

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