The full post is in this pdf file with R code for simulation and empirical examples:
In this post, I explain the problems and deficiencies associated with unit root testing. As a result of these deficiencies, the unit root test has generated so many puzzles and controversies since it was first introduced in early 1980's to economics.
It is quite likely that most of these results are the outcomes of Type II errors, where the false null hypothesis of a unit root is accepted due to a range of problems including the low power.
I discuss the unit root test, its implications in economic time series, its problems, and how its outcomes have generated a range of puzzles and controversies. The U.S. real GDP data is used as an example.
Comments