Test for a structural break with an unknown break date / postestimation cannot continue due to collinear variables

[Marcel Wieting]

Hi all,

I am running the sbsingle command to test for a structural break with unknown break date.

reg y L1.y x z
estat sbsingle

I get the following error code:

postestimation cannot continue due to collinear variables
r(198);

However, if I check the VIF scores on all variables implemented into the regression, there is no indication of colinearity.
What am I doing wrong? I am using STATA 15.1
Best,

Marcel

[Phil Bromiley]

You didn't get a quick answer. You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output and sample data using dataex.

Generally, you can't include a lagged dv in a model without controlling for the strong likelihood it is endogenous. see xtivreg, xtivreg2, xtabond, xtdpdml. This is discussed explicitly in the estat sbsingle documentation.

estat sbsingle appears in the time series documentation. If may only work for a single time series. You might try the documentation's suggestion on trimming.

[Jet Vollers]

Dear Marcel Wieting,

Did you solve this problem already? I'm experiencing the exact same thing.

Regards,
Jet