Bookmark and Share

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


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

Re: st: Multicollinearity warning with condivreg


From   Maarten buis <[email protected]>
To   [email protected]
Subject   Re: st: Multicollinearity warning with condivreg
Date   Mon, 12 Apr 2010 06:34:17 +0000 (GMT)

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany

http://www.maartenbuis.nl
--------------------------


--- On Mon, 12/4/10, Dana Shills wrote:
> I am estimating an instrumental variable regression with a
> single endogenous regressor and a single instrument. (Among
> other control variables are also several dummies for
> industry, country, and time). The first stage F-stats are
> very high and I am interested in reporting the conditional
> likelihood ratio confidence intervals.
>  
> When I use the condivreg command with the Moreira and Poi
> update (st0033_1 in SJ 5(2)), the estimation runs through
> without providing any warnings (however the output in this
> version doesn't provide confidence intervals). When I update
> the command using the more recent update by Mikusheva and
> Poi (st0033_2 in SJ 6(3)), I get "Multicollinearity"
> warning. (I check both the first and second stage
> regressions separately using OLS and VIFs are all less than
> 10). What am I doing wrong? 

Have you for each categorical variable (industry, country and
time) a reference category? So, say industry consists of the
categories agriculture, mining, fishing, industry, and 
services, then you would need to leave one of the dummies out
of your model (e.g. the dummy for agriculture).

Hope this helps,
Maarten


      

*
*   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/


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