Leader value added: Assessing the growth contribution of individual national leaders

Previous literature suggests that leaders matter for growth in general. This paper asks: which leaders matter? We develop a method to estimate the idiosyncratic growth contribution of individual leaders (relative to their compatriots) and, drawing on Armstrong et al. (2022), calculate its uncertainty via a robust empirical Bayes confidence internal (EBCI). We show that most intuitive estimate of a leader's contribution—the average growth rate during their tenure—is mostly useless for measuring his or her contribution and estimating the leader effect requires major adjustments to that raw growth average. Relatively few leaders have large contributions with zero outside the 95% robust EBCI, which suggests it is difficult to identify which leaders are good or bad for growth using growth data. Although there are some leaders that stand out using our method, many of those leaders are surprises. Moreover, leaders in non-democratic countries are not much more likely to be identified as good or bad for growth than leaders in democratic ones.