"Radford Neal" <[EMAIL PROTECTED]> a �crit dans le message news:
> In any case, the original poster explicitly claimed that regression
> with an explanatory variable that was generated by a non-stationary
> process was invalid even if the residuals of the regression are
> independent. I claim that this is not true.
>
Since i am not an expert in cointegration, I cannot prove it, sorry.
However, here is a quote from A. Banerjee et al.,"Cointegration,
error-correction, and the econometric analysis of non stationary data",
Oxford university press, p.167 :
"The importance of the later point follows from the observation that, even
when the regressand (e.g. y(t)) and the regressor (e.g. x(t)) are both
integrated of order one and are cointegrated, the t-statistics on the
coefficient of
x(t) still has a non standard distribution which makes ordinary t and normal
tables unusable for purposes of inferences"
The regression equation with iid errors implies cointegration of the two
series.
I am not certain of it, but i think that the mean of the mean is OK, but the
variance of the estimator of the mean is not (which could explain why t-test
won't work)
David B
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
