The article discusses a new way to test if one thing causes another. Instead of using tests that look at past data, the researchers created tests that predict future outcomes. They found that using these new tests can change the results and conclusions about how money affects output.