This paper provides evidence on the unit root hypothesis and long-term growth by allowing for two structural breaks. We reject the unit root hypothesis for three-quarters of the countries - approximately 50% more rejections than in models that allow for only one break. While about half of the countries exhibit slowdowns following their postwar breaks, the others have grown along paths that have become steeper over the past 120 years. The majority of the countries, including most of the slowdown countries, exhibit faster growth after their second breaks than during the decades preceding their first breaks.