Structural Break Tests with an unknown date - Ambiguous information given in STATA output.

[Thomas Obst]

Hello,

I am running a structural break test with an unknown data (command in STATA: estat sbsingle). In the attached file you see the output I receive. As you can see, the null hypothesis is not rejected indicating that there is no structural break. However, the output also reports and estimated break date (2007q4). Hence, I am wondering whether there is a structural break at 2007q4 or not? Can you please explain how to decide in this case whether there is indeed a structural break or not?

Best regards
Thomas

[Phil Bromiley]

It gives you an estimate for structural break, but cannot reject the hypothesis that there is no break. Just like parameters in a regression appear even if they're not statistically significant.
By the way, can use L.y instead of creating the lag before the regression. Also, often regression with a lagged dv is not consistent which is why there are a bunch of dynamic panel estimators.

[Thomas Obst]

Dear Phil,
thank you for the responds. This is what I suspected. So in fact there is no structural break, and STATA simply reports the estimate. I will take on board your suggestion with L.y. Can you please elaborate a bit more on your point of using a lagged dependent variable. Why is it not consistent? Consistent with what exactly?

Thank you.

Best regards
Thomas