Re: st: AKAIKE formula

[Maarten buis <maartenbuis@yahoo.co.uk>]

--- On Tue, 13/4/10, Paulo Regis wrote:
> Lets say it is an F-test, where one of the "q" 
> restrictions is the coefficient of the lag of the dependent
> variable. In my example, I showed the case where q=1. 
> H0: c = 0 .
> 
> More general for an F-statistic, since you have the lag of
> the dependent variable, there is endogeneity (you can use
> lags of the indep variables as instruments). Therefore, you
> can think of a Wald test where:
> a) you use OLS for the restricted model.
> b) you should use IV for the unrestricted model.
> 
> Alternatively, I could check which of the two models
> minimize akaike.
> How inappropriate is that?

I am guessing you mean the following:

- you have model where you believe that you need to use an
IV estimator because you added a laged dependent variable
as a predictor.

- You have a second model that did not include the laged
dependent variable, so you concluded that you did not
have to use an IV estimator

Both models contain a variable x

You are interested in the hypothesis that the effect of x
is equal in both models.

Sounds a bit similar to a setup for a Hausman test, so maybe 
you can look into that literature to see if that can give
you some inspiration on how to takle your problem.

Maybe somebody else on the list has a bright idea?

-- Maarten

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany

http://www.maartenbuis.nl
--------------------------


      

*
*   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/