"Asiedu, Elizabeth" wrote:
>
> Dear all,
>
> I have an unbalanced panel of 91 countries over the period 1984-2000 (1024
> observations) and I'm using the fixed effects estimation, because it failed
> the Hausman test. I have two questions:
>
> 1. How do I test whether there is endogeneity?
>
> 2. I can't reliable instruments for my endogenous variable. How do I resolve
> the problem of endogeneity if it exists?
As the previous posters indicated, the usual way of testing for endogeneity
involves an instrument (which you don't have). But Wooldridge, Econometric
Analysis of Cross Section and Panel Data, 2002, p. 285 suggests a test for
strict exogeneity that might suit your need. The test is implemented by
estimating the equation including *leading* values of the variable(s) that you
suspect might not satisfy the assumption of strict exogeneity. With respect to
the second question, look at section 11.1.2 of the same book for an explanation
of the circumstances in which twice-lagged values of regressors could be used as
instruments in a first difference model.
Dave Harless
*
* 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/
