dfgls unit root test regression output?

[Rene Macon]

Is there a possiblity to see the regression estimates of a "dfgls" unit root test in a similar way as it is possible to see the "dfuller" unit root test regression estimates (e.g. adding an option like "regress")?
Thanks a lot for any help!
Rene

[Marcos Almeida]

Welcome to the Stata Forum / Statalist,

I'm not sure if I understood the reason of your demand. That said, to my knowledge, there is not option "regress" for the - dfgls - command.

[Rene Macon]

That's right, the option "regress" does not work for the "dfgls" test. Nevertheless, I would like to know whether the estimated trend coefficients are significantly different from zero...
Thanks a lot for a feedback!
Rene

[Marcos Almeida]

Rene Macon

Nevertheless, I would like to know whether the estimated trend coefficients are significantly different from zero...

With regards to the null hypothesis of dfgls test, I gather this explanation from the Stata Manual will clarify your query:

dfgls performs the DF-GLS test for the series of models that include 1 to k lags of the firstdifferenced, detrended variable, where k can be set by the user [...] . As discussed in [TS] dfuller, the augmented Dickey–Fuller test involves fitting a regression [...]and then testing the null hypothesis H0: β = 0. The DF-GLS test is performed analogously but on GLS-detrended data. The null hypothesis of the test is that yt is a random walk, possibly with drift. There are two possible alternative hypotheses: yt is stationary about a linear time trend or yt is stationary with a possibly nonzero mean but with no linear time trend. The default is to use the former. To specify the latter alternative, use the notrend option.

Hopefully that helps.

[Rene Macon]

Thanks a lot Marcos! I've actually tried the alternative null hypotheses, explained in the Stata Manual. According to the theory, I want to test, my yt should be stationary around a nonzero mean. And indeed, I cannot reject this version of the null hypothesis. Now, I would like to know, wether the estimated mean is actually "nonzero", because the null leaves this open ("with a possibly nonzero mean"). Do I make a mistake? Is this a wrong question? I'd be grateful for a feedback!
Best wishes,
Rene

[Marcos Almeida]

There is not such a thing as the ' alternative null hypothesis '.

Indeed, there is only one 'null hypothesis'. In this case , as it seems, the null hypothesis was not rejected, hence yt is considered as a random walk.