Sequential break testing with unit roots

Develop the asymptotic distribution theory for sequential tests (such as sup-F or sup-Wald procedures) that estimate the number and locations of structural breaks in linear time series regression models when unit roots are present.

Background

While wild-bootstrap procedures have been validated for single-break tests in the presence of unit roots, a full asymptotic theory for sequential multiple-break testing with unit roots has not been established. This limits practitioners’ ability to conduct rigorous inference on the number of breaks under unit-root processes.