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: Re: st: Why do i get different results with year robust clusters vs year dummies


From   Haillie Lee <[email protected]>
To   [email protected]
Subject   Re: Re: st: Why do i get different results with year robust clusters vs year dummies
Date   Tue, 3 May 2011 10:46:27 -0400

Thank you very much for your help. My data set covers 188 countries
and 64 quarters. I am using year fixed effect instead of country fixed
effect because my estimation controls for the most of conventional
contry variables -- which makes me worried more about the quaterly
variations. Plus, I thought since I have more countries than quarters,
it would make more sense to cluster on countries rather than years.

I am still a novice in the stats world so any advice you can provide
would be greatly appreciated. Thank you very much!

Sincerely,
Haillie

On Tue, May 3, 2011 at 7:11 AM, Christopher Baum <[email protected]> wrote:
> <>
> On May 3, 2011, at 2:33 AM, Haillie wrote:
>
>> Thank you very much for your reponse. In other words, does that mean
>> that if i do robust cluster(country) absorb(year) it would be
>> equivalent to introducing year fixed effects? When i tried this, I
>> still found my coefficient to be different compared to the year
>> dummies model. However, this could be due to the fact that some of the
>> year dummies were omitted due to collinearity problem. Could this be
>> the reason?
>
> I don't use -areg- (instead, I use -xtreg, fe-), but my understanding is that -areg- with absorb() is the equivalent of -xtreg, fe- where you have declared the panel variable as that which is being absorbed (perhaps with xtset panelvar). Both of those should then be equivalent to estimating the model with i.year as a regressor. But -xtreg, fe- would insist that you cluster on the panelvar, and not on some other variable such as country. What are the dimensions of your data in terms of country and time? If you have fewer than ~40 countries and/or time periods, it is probably not a good idea to use cluster-robust VCE in that direction.  But don't you expect unobserved heterogeneity to be a bigger issue in the country dimension than the time (macro) dimension? Why aren't you using country fixed effects?
>
> Kit
>
> Kit Baum   |   Boston College Economics & DIW Berlin   |   http://ideas.repec.org/e/pba1.html
>                              An Introduction to Stata Programming  |   http://www.stata-press.com/books/isp.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/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/
>



-- 
Haillie Lee
PhD Student
Department of Government
Georgetown University

*
*   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