st: Re: Stata Time Series testparm question

[Michael Hanson <mshanson@mac.com>]

On Jul 17, 2007, at 5:22 PM, Laura Clavijo Muñoz wrote:

> Dear Mike, I was reading your thread which I post below:
>
>
> > “Perhaps you have a "Granger causality" test in mind? This can be  
> > accomplished in Stata after estimating a -var-, but it really is  
> > just a test of joint significance of the lags of x. That is, for  
> > example,
> >
> > var y x, lags(1/3) small dfk
> > vargranger
> >
> > and
> >
> > reg y L(1/3).y L(1/3).x
> > testparm L(1/3).x
> >
> > should give you exactly the same answer.”
> In trying to do the same, Stata doesn´t let me use time-series  
> operators with testparm. Because I´m using a loop to choose the  
> optimal lags, I need to be able to use this format L(1/x). Any idea  
> what is happening?
>
> Thank you very much.
	The convention on Statalist is to respond to messages posted there  
rather than to individuals.  That way, many people potentially can  
help you, and any help you receive becomes a public good.  Hence, I  
have responded to Statalist in my reply.  If you have not yet joined  
Statalist (i.e., you quoted my message from a web-based search), I  
encourage you to do so.  See <http://www.stata.com/support/faqs/res/ 
statalist.html> for information about subscribing.

	With regard to your question, please consider the following set of  
commands that generalize my quoted sample code.  You should confirm  
that these commands work for you as well (i.e., cut and paste them  
into a new .do file, and run that in your copy of Stata).

// Begin example code

webuse lutkepohl
gen y = lincome
gen x = lconsumption

set more off
forval q = 1/4 {

	var y x, lags(1/`q') small dfk
	vargranger

	reg y L(1/`q').y L(1/`q').x
	testparm L(1/`q').x

}

// End example code

	One possibility is that your reference to the looping variable  
(which you refer to as "x" above) is not understood by Stata.   
Sometimes a right tick (') inadvertently is substituted for a left  
tick (`) -- a good text editor with syntax highlighting is helpful in  
such cases.  Otherwise, the admonition on the Statalist FAQ would  
seem to apply here:  "Messages to the effect that some command did  
not work for you are often impossible to answer without seeing  
precisely what you typed and what Stata returned in response."

	Another possibility is that you are using a different (possibly  
older) version of Stata.  These commands work with version 9.2 (which  
I just confirmed);  the latest version of Stata is 10.  You may want  
to see if you get similar versions and dates from the following three  
commands as I do;  if one or more is different, that might point  
towards the source of your difficulties.  (And note also from the  
Statalist FAQ:  "Please specify if you are using an earlier version;  
otherwise, the answer to your question is likely to refer to commands  
or features unavailable to you.")


. which testparm
/Applications/Stata/ado/base/t/testparm.ado
*! version 3.1.3  27jul2004

. version
version 9.2

. update query
(contacting http://www.stata.com)

Stata executable
    folder:               /Applications/Stata/
    name of file:         StataSE
    currently installed:  12 Jan 2007
    latest available:     12 Jan 2007

Ado-file updates
    folder:               /Applications/Stata/ado/updates/
    names of files:       (various)
    currently installed:  22 May 2007
    latest available:     22 May 2007

Recommendation
    Do nothing; all files up to date.


	Hope that helps.

                                        -- Mike


*
*   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/