Michael Paris schrieb:
Hi
I have the below results. I would like to make sure I read correctly
Test with constant and time trend
T=27
hypothesis of unit root : a=1
model : (1-L) y = b0 + b1*t + (a-1) * y(-1) + ... + e
Coeff. autocorrelation of 1st order for e : -0.012
Estimated value of (a-1) : -0.65
Test statistic : tau_ct (1) = - 3.05
p. asymptotic critic : 0.11
=> For a probability of 0.11 and n=27, the test statistic should be <
-3.18 as per df distribution tables in order to be stationary. So this
series is not stationary. Is this correct?
Thanks a lot for the help!
Mike
------------------------------------------------------------------------
Windows 7: Find the right PC for you. Learn more.
<
http://windows.microsoft.com/shop>
------------------------------------------------------------------------
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users Hey Mike,
It just means that you cannot reject the null of non-stationarity under
the alternative of a stationary series not even at the 10% level. Thus,
the series is assumed to be non-stationary (random walk).
Best,
Artur