st: re: predicting y, with other variables

[Kit Baum <baum@bc.edu>]

sacrificial radish

Alessia said

I try to explain it better then. I didi my regressione, kept my b. Now
there is a x variable in which I did a variation (I took one of its
observation and I modify it, saying doing a 1% variation. I want then
to know how the predicted values are going to change).

regress y x1 x2
mfx compute, dyex

will compute the effect on the level of y of a 1% change in each x.


Re other q's:

1) you can always compute an R^2-like measure from the squared simple  
correlation of Y and Yhat.

2) estimating a model with difference GMM is an instrumental  
variables estimator. One would expect any IV estimator to give you  
different results than would OLS on levels if your maintained  
hypothesis in running difference GMM -- that some of the explanatory  
variables are correlated with the error -- is appropriate.

3) See the Baum-Schaffer-Stillman paper (Stata Journal, 2003) or the  
working paper version available from my RePEc CV below. If you're  
going to use xtabond2, read David Roodman's "how to do  
xtabond2" (available from IDEAS or EconPapers).



Kit Baum, Boston College Economics and DIW Berlin
http://ideas.repec.org/e/pba1.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/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/