st: Re: Newey estimations

[Kit Baum <baum@bc.edu>]

Thanks to Clive for his enthusiastic endorsement of the "wildly  
popular"* ivreg2 and reminder that yes, it does OLS, HOLS and "Newey- 
West" if asked nicely (although a somewhat less bellicose set of  
references to its strengths might be appropriate given the daily  
contents of CNN).

To clarify, ivreg  bw(2) == newey lag(1)
                             bw(5) == newey lag(4)
and so on.

ivreg2 also implements numerous other kernel estimators (beyond the  
Bartlett kernel employed by newey) and it may be worth trying them  
out. There is nothing sacrosant about Bartlett; other kernels may do  
a better job of estimating the LR variance in finite samples.

A followup paper to Baum, Schaffer, Stillman, Stata Journal 3:1 2003  
describing all of the new bells and whistles in the ivreg2 suite is  
now in draft form. When we finalize the draft and submit to SJ, we  
will circulate the draft.

* That's a quote from StataCorp, not from the authors.

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


On Aug 3, 2006, at 2:33 AM, Clive wrote:


> All of which is to say, download -ivreg2- from SSC. The -bw(2)-  
> option is
> Newey-West.
*
*   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/