A robust bootstrap test under heteroskedasticity

A test for an unknown structural break based on an heteroskedasticity-robust artificial regression is considered. The test has good finite sample properties under different resampling procedures and transformations of the residuals.