Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: RE: time dummy or trend


From   "Neuling,MM (pgt)" <M.M.Neuling@lse.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: RE: time dummy or trend
Date   Wed, 10 Aug 2005 21:55:55 +0100

the trend-stationary time series that i have is in log-first-difference... i checked the process graphically and just added a trend line: both linear as well as power two ... both seem to fit well... i dont know whether this is a good indication ... however, i think what i generated is a time line counting up in integers (1 2 3...)  and hence a variable that is a linear process which should capture the trend in the dependent variable ... i ll have another look into it though and see what i can still do... thanks a lot for your help ... take care, matthias 

	-----Original Message----- 
	From: Nick Cox [mailto:n.j.cox@durham.ac.uk] 
	Sent: Wed 8/10/2005 7:01 PM 
	To: statalist@hsphsun2.harvard.edu 
	Cc: 
	Subject: st: RE: time dummy or trend
	
	

	This seems more a matter of statistical logic
	than one of Stata syntax.
	
	Given that you have a time variable, there seems
	little point in coarsening it to quarters.
	
	Also, the -int()- here presumably does
	nothing.
	
	A regression with -date- as predictor
	would be fitting a linear trend
	component in that variable; a regression
	with -time- one that would give
	similar (but possibly better)
	performance. The b coefficient would
	differ because of the differing units.
	
	Keeping track of things graphically
	would enhance your understanding of
	what you are doing.
	
	Nick
	n.j.cox@durham.ac.uk
	
	Neuling,MM (pgt)
	
	> i have little experience with stata8 and hope one of you may
	> help me...
	> 
	> i have a time series regression with several explanatory
	> variables, the dependent variable being trend-stationary... i
	> have generated the following variable to account for the
	> trend in the process and to inlcude it in the regression:
	>
	> gen date = quarterly(time, "yq")
	>
	> tsset date, quarterly
	>
	> gen time_ind= int(date)
	>
	> when i run the regression with the "time_ind" included as
	> explanatory variable, does it take care of the trend and is
	> it a trend or time dummy?
	
	*
	*   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/
	

<<winmail.dat>>




© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index