We investigate how the length of time a country's regime was autocratic between 1920 and 2000 is correlated with economic growth and per capita income. We find that the longer a country was within an autocracy, the lower is the country's economic performance, even after controlling for other factors. We also find the length of time a country is not autocratic is positively related to growth and income. We claim this evidence is consistent with the thesis that one reason why some countries have had difficulty adjusting to life after autocracy is that the human and social capital necessary to make markets “work” eroded under autocratic regimes and take time to develop afterward. (JEL O17, O43, P0)