The article discusses new ways to test for unit roots in panels with cross-sectional dependence. The researchers introduce block bootstrap versions of two popular unit root tests for panel data. These tests can be easily applied without needing to know the specific dependence structure. They show that the bootstrap tests have better performance in simulations compared to traditional tests, especially in settings with common factors or cointegration. This means the new tests are more reliable for detecting unit roots in real-world data.