Statalist The Stata Listserver


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

st: instrumental variables


From   Kit Baum <baum@bc.edu>
To   statalist@hsphsun2.harvard.edu
Subject   st: instrumental variables
Date   Sun, 19 Nov 2006 10:59:58 -0500

George said

i am running instrumental variables estimations. y=f(X,W), and W is endogenours, so stage one becomes W=f(Z), and stage 2 becomes y=f (X, Predicted W). My function is aproduction function with conventional inputs such as fertiliser, seeds, labour and then credit which is the (W). Must the number of RHS variables in stage one be the same number as RHS ones in stage two.

George should heed the Statalist FAQ and use informative subject headings on his postings.

He should read any textbook discussion of instrumental variables / two stage least squares. He might also find the Baum/Schaffer/ Stillman Stata Journal article useful; a working paper version can be accessed from
http://ideas.repec.org/a/tsj/stataj/v3y2003i1p1-31.html

One does not run the first stage and second stage as separate regressions. The number of included endogenous variables (one in his example, W) must be less than or equal to the number of excluded instruments. His mention of Z does not make clear whether that is a variable or a matrix of variables. If Z is a single variable, the equation would be exactly identified via the order condition. It would be preferable if the equation were overidentified---if Z contained more than one column---for then Sargan/Hansen tests for overidentifying restrictions could shed some light on the adequacy of the instruments.

Kit Baum, Boston College Economics
http://ideas.repec.org/e/pba1.html
An Introduction to Modern Econometrics Using Stata:
http://www.stata-press.com/books/imeus.html


*
* 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/




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