slope dummy

[Meg]

Hi
I am trying to find the trend of a series; Y.
The series have at least one structural break. I used a dummy variable and regressed Y over @trend, with and without the dummy. Am I going in a right way to use dummy variable to find the trend? The trend without dummy is just like simple regression.
Should I also use seasonal dummies?
Many thanks
Meg

[Meg]

Based on unit root test results Y doesn't have a trend.
But my question is that if there is structural break are the unit root tests reliable?
Many thanks
Meg

[startz]

Unit root tests only check for a stochastic trend, not a deterministic trend. And no, if there is a structural break the usual tests aren't right.

[Meg]

Thanks dear startz for your reply and defining that unit rot tests are not valid in presence of structural break.

Could you please advise on which method I should use to find the trend for Y, the file was attached to my first post.

regards
Meg

[startz]

To test for a unit root with one structural break, start with the Zivot-Andrews test. Identifying trends is a big topic, with no single, simple answer.