National leaders ¨C especially autocratic ones ¨C are often given credit for high average rates of economic growth while they are in office (and draw criticism for poor growth rates). Drawing on the literature assessing the performance of school teachers and a simple variance components model, we develop a new methodology to produce optimal (least squares) estimates of each leader¡¯s contribution to economic growth (controlling for commodity prices, regional business cycles, and country effects). While we do sometimes find sizable growth contributions of celebrated ¡°benevolent autocrats¡±, we also find that (i) they are regularly outranked by other less celebrated leaders and (ii) the ranking and contributions of leaders is often not robust across growth datasets. Moreover, we find that even in world where leaders do affect growth, the average growth rate during a leader¡¯s tenure is mostly uninformative about that leader¡¯s actual growth contribution. Depending on the dataset and methodology, we find that that measured least squares leader contributions and unobserved leader effects can vary just as much in democracies as autocracies.