Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: GMM estimation and Helmert procedure

From   Abubakr Saeed <>
To   "''" <>
Subject   st: GMM estimation and Helmert procedure
Date   Sat, 25 Feb 2012 13:47:09 +0000

Dear Statalist users,

I have an unbalanced panel dataset and intend to running investment equation with GMM. I have two questions regarding this:

First, there are few studies that have used forward-mean differencing procedure, also referred to as the Helmert's procedure, to removed the fixed effects. But my question is when we are estimating GMM in first-differences do we still need to transform our variables through forward-mean differencing procedure (because fixed firm-specific effects are already removed by estimating GMM in first-differences)?  If I have to use the forward-mean differencing procedure then how can I do it? I tried to find the Helmert procedure but couldn't get it.

My second question is bit related to syntax of GMM-FD. 

xtabond2 depvar l.depvar l.indepvar1 l.indepvar2 l.indepvar3 if country==xxx, gmm (l.depvar l.indepvar1 l.indepvar2 l.indepvar3, lag (2 2)) iv(l2.indepvar1) nolevel small

I am assuming my all variables are endogenous, having 9 years time period. Now since I don't have any additional IV variables that I can treat sickly exogenous, so is it possible to not provide any IV,..... if not then how can I use same endogenous variables here as exogenous as well. I addition, somebody suggested me to use l2.indepvar1 as IV but I have already instructed Stata to use only the second lag of the endogenous variables. I am really blocked here, and any suggestion would be helpful.

Thanks so much for your insights.


Abubakr Saeed
PhD student

*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index