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

st: xtabond(2) and IV


From   Kit Baum <[email protected]>
To   [email protected]
Subject   st: xtabond(2) and IV
Date   Thu, 9 Dec 2004 14:10:10 -0500

Danielle wrote

Can someone please clarify what Stata does w/ the following two commands?

1) xtabond2 sex_12 L(sex_12 binge_30), gmm(L(sex_12 binge_30)) noleveleq
2) xtabond2 sex_12 L(sex_12 binge_30) age_mo, gmm(L.sex_12) noleveleq

Actually, I don't really want to do this, but in the interest of
understanding how <xtabond2> and <xtabond> work, suppose I want to
instrument for lagged diff of sex_12 using lagged values of sex_12 &
lagged values of binge_30 but I do not want to instrument for lagged
diff of binge_30. Will (1) get this done? (binge_30 is endog but the
lagged diff of binge_30 does not need to be instrumented for b/c it is
not correlated w/ the error term)

Suppose I want to instrument for only lagged diff of sex_12 using only
lagged values of sex_12. I want to treat lagged diff of binge_30 as
exog. Does (2) do this?


As a prior posting on this thread and many previous postings by Mark Schaffer and I have indicated, an instrumental variables estimator (and IV is a special case of GMM) does not work by instrumenting X1 with Z1, X2 with Z2, etc. All instruments are in the Z matrix, and all regressors are in the X matrix. The fact that some Zs are also Xs is taken for granted. Things get more complicated when the A-B estimator is considered, but the same issue applies: you can't designate instruments to specific regressors.

Kit Baum, Boston College Economics
http://ideas.repec.org/e/pba1.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 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index