st: re: Consequences of OLS with poisson distribution

[Christopher Baum <baum@bc.edu>]

<>
Ronggui said
Besides the facts that prediction of ols can be less than zero which  
is impossible in Poisson distribution. I also know that the standard  
error is inconsistent. But I cannot find reference on what is the  
exact meaning of inconsistent. That is, is it upwards, downwards or it  
depends?

As any econometrics text makes clear, an estimator is consistent if  
its estimate converges upon the true (population) value of the  
parameter as sample size goes to infinity: that is, it is a large- 
sample property of the estimator. Thus
xtilde= sum(x)/(N-1)
is a biased but consistent estimator of sample mean, while
xttheta = 2/3 sum(x)
is not. With regard to the standard errors of estimated parameters, we  
cannot in general say whether an inconsistent estimate of the variance- 
covariance matrix (VCE) delivers s.e.'s that are too large or too  
small (except in contrived cases).
Kit
Kit Baum   |   Boston College Economics and 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/