EViews.com

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

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

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.

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

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.