How can I calculate Newey-West standard errors in an
IV regression? As an example, say I want to estimate
the model:
y=x1 + x2 + x3 + e,
where x3 is endogenous and Z is my instrument so that,
x3=x1+x2+Z+u will be the first-stage equation.
To estimate in Stata, I can run
>ivreg y x1 x2 (x3=z)
It can be run separately too:
>regress x3 x1 x2 z
>predict x3_hat
>regress y x1 x2 x3_hat
The coefficients on x1, x2, and x3_hat will be the
same in these steps as in the -ivreg- command but the
standard errors will be wrong (secondary question: how
would one do this and recover the correct standard
errors?).
To get the Newey-West standard errors for the second
stage, I assume I can't just use the -newey- command
after the last -regress- command. So, how can I adjust
the second-stage standard errors for autocorrelation?
Thanks,
Jon Schwabish
Partnership for New York City
[email protected]
__________________________________
Do you Yahoo!?
Yahoo! Mail - Find what you need with new enhanced search.
http://info.mail.yahoo.com/mail_250
*
* 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/
