st: multilevel model--endogeneity

[Steven Archambault <archstevej@gmail.com>]

Hi all,

I am analyzing household consumption data that contains information at
the village and district level for each household using a multilevel
model. I am suspecting that one of the district level explanatory
variables has an endogeneity issue with one or more of the village and
household level variables. I read a paper by Spence and Fielding (An
Instrumental Variable Consistent Estimation Procedure to Overcome the
Problem of Endogenous Variables in Multilevel Models) where a first
stage multilevel regression is carried out to analyze the variable
with the endogeneity problem, using exogenous instruments and the
other explanatory variables from the primary model. The predicted
values are them plugged back into the primary model, as the second
stage.

Does this seem appropriate for GLLAMM, or maybe xtmixed? How would
others take care of this issue?

Thanks!
-Steve
*
*   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/