As written, p is not endogenous in eqn (2), but it should be, since p
and q are determined by the intersection of supply and demand curves
both shifted by various factors over time, and no, -xtreg- is not
sufficient--you should find an instrument and use -xtivreg2- (
ssc inst xtivreg2
ssc inst ivreg2
ssc inst ranktest
etc.
) to estimate eqn (2). For details, read the help file for -ivreg2-
and its refs.
As for reposting the same message with a different subject heading,
http://www.stata.com/support/faqs/res/statalist.html#noanswer
offers different advice, and if you are concerned about the lag
between sending a msg to Statalist and its showing up on the archive,
check your headers which tell you that the 6-hour lag is in your(?)
mail server:
Received: from fmmailgate09.web.de (fmmailgate09.web.de
[217.72.192.184]) by guardian.sph.harvard.edu with ESMTP id
6fPrIUsgNUURXnHl for <[email protected]>; Thu, 04 Sep
2008 05:53:20 -0400 (EDT)
Received: from web.de
by fmmailgate09.web.de (Postfix) with SMTP id 97A342A4F97D
for <[email protected]>; Thu, 4 Sep 2008 11:53:16 +0200 (CEST)
On Thu, Sep 4, 2008 at 5:48 AM, <[email protected]> wrote:
> Hello everyone,
>
> I have a small N, large T panel data set and I would like to estimate the price elasticity of demand.
>
> My system reads:
>
> (1) p(i,t) = a0(i) + a1*p(i,t-1) + u(i,t)
> (2) q(i,t) = b0(i) + b1*p(i,t) + v(i,t)
>
> Hence, my question is: Which estimation procedure is appropriate, given that p(i,t) is endogeneous in equation (2) and depends on its own lag in equation (1)? Furthermore, are standard panel techniques applicable - my panel is NOT large N, small T but vice versa?
>
> I'm only interested in the parameter b1 (variables are in logs, thus b1 is the price elasticity of demand). Can I estimate (2) without respect to the endogeneity of p(i,t) with say xtreg?
>
> Many thanks in advance for your comments!
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
