The article introduces a new method to detect changes in relationships between variables over time. By using a modified adaptive lasso estimator, the researchers can accurately identify structural breaks in cointegrating regressions. They also develop tests to determine if these breaks affect the long-term relationship between variables, showing that their tests have the power to detect cointegration even in the presence of multiple structural breaks. The findings suggest that the timing and magnitude of these breaks can be consistently estimated, providing valuable insights into the dynamics of these relationships.