Dear all,
Just wanted to ask a simple clarification. When running the Breitung panel unit root using xtunitroot, is the alternative hypothesis homogeneous (that all panels are stationary, as in LLC test) or heterogeneous (at least one panel is stationary, as in the IPS test). Both the stata output and manual seem to indicate homogeneous, but my results are more similar to the IPS test and some sources indicate the Breitung test can be applied to the heterogeneous alternative (e.g "The DGP is the same as in the Im, Pesaran, and Shin approach" at the following link https://documentation.sas.com/doc/en..._details50.htm).
So I just wanted to check that Stata indeed uses the homogeneous alternative by default (indeed, I can't see anywhere in the manual where it indicates specifying which alternative hypothesis to use.
Thanks!
Just wanted to ask a simple clarification. When running the Breitung panel unit root using xtunitroot, is the alternative hypothesis homogeneous (that all panels are stationary, as in LLC test) or heterogeneous (at least one panel is stationary, as in the IPS test). Both the stata output and manual seem to indicate homogeneous, but my results are more similar to the IPS test and some sources indicate the Breitung test can be applied to the heterogeneous alternative (e.g "The DGP is the same as in the Im, Pesaran, and Shin approach" at the following link https://documentation.sas.com/doc/en..._details50.htm).
So I just wanted to check that Stata indeed uses the homogeneous alternative by default (indeed, I can't see anywhere in the manual where it indicates specifying which alternative hypothesis to use.
Thanks!
