[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: Prediction after xtpoisson, fe,

From   Pierre Azoulay <>
Subject   st: Prediction after xtpoisson, fe,
Date   Wed, 22 Jul 2009 17:36:44 -0400

* if you are using xtpoisson, fe, you are NOT following the
recommendation of Santos Silva and Tenreyro. These will produce ML
standard errors. You want QML "robust" standard errors. Unfortunately,
xtpoisson does not offer this option -- a shame that stata 11 does not
appear to remedy.
Tim Simcoe from Toronto/BU has developped an ado file that modifies
xtpoisson to do this.

* With QML, I am not sure it is valid to predict ion that way, since
the only assumption you are making is on the conditional mean.

hope that helps,


Pierre Azoulay
Associate Professor
Massachusetts Institute of Technology
Sloan School of Management
50 Memorial Drive — E52-555
Cambridge, MA 02142-1947

Tel [Sloan]: (617) 258-9766
Tel [NBER]: (617) 588-1464
Fax: (617) 253-2660

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index