[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Interdependent panel system

From   "Austin Nichols" <>
Subject   Re: st: Interdependent panel system
Date   Thu, 4 Sep 2008 10:54:33 -0400

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

) 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,
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 (
[]) by with ESMTP id
6fPrIUsgNUURXnHl for <>; Thu, 04 Sep
2008 05:53:20 -0400 (EDT)
Received: from
	by (Postfix) with SMTP id 97A342A4F97D
	for <>; Thu,  4 Sep 2008 11:53:16 +0200 (CEST)

On Thu, Sep 4, 2008 at 5:48 AM,  <> 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:

© Copyright 1996–2015 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index