Looking for Leadership in Historical Context: An Extension of the RIFLE Method of Randomization Inference

A growing literature supports the idea that individual political leaders can have a profound impact on economic and other outcomes. We extend Berry and Fowler's (2021) RIFLE technique to pre-modern historical contexts. First, we demonstrate the technique's applicability to long-duration historical data of the type commonly found in historical research. We revalidate the technique on simulated data to show it performs well in this environment. Second, we illustrate RIFLE's potential using several applications that point to where and when pre-modern leaders exerted influence over outcomes. Specifically, we test the leadership effect of European and English monarchs and ministers on economic growth in the pre-modern period and find no evidence of substantial leadership effects. Finally, we test for leadership effects of Roman emperors on the contracting and expanding borders of the empire and find robust evidence of a leadership role, especially in defense. These results are a first step towards validating the role of leadership in historical outcomes. Many additional avenues for research remain, with interesting new methods for testing the role of leader agency in disparate political contexts.