The article compares different tests for checking if economic data has a unit root. Panel tests are better at this than single tests. Moon and Perron's tests work well for different data scenarios. Bai and Ng's ADF test lacks power for common factors, but the pooled Dickey-Fuller-GLS test is better for individual components. Choi's tests are too big when common factors affect units differently. All tests struggle with trends in the data.