The Power of "Big Brothers": Political Leaders' Impact on Authoritarian Regimes

• In Event: Autocratic and Traditional Leaders

Abstract

Why do some authoritarian regimes consolidate while others liberalize? Comparative political scientists have proposed an array of explanations, including factors such as economic performance, political institutions, state capacity, and political culture. Remarkably, the role of political leaders has received limited attention in large-N studies. This is a regrettable gap since, for dictatorships, the dictators are “elephants in the room.” The collapse of authoritarianism in the Soviet Union around 1989 is likely attributed to stagnant economy or ossified political institutions, but “the Gorbachev factor” is undeniably striking (Brown, 1997). Had Chernenko lived for another decade, perhaps the year 1989 would not have left such a special mark on history. There were probably many reasons why China escaped the democratization wave in 1989, but Deng’s decision to suppress the student movement was the pivotal turning point. If Deng had passed away ten years earlier, the future trajectory of China would remain speculative.

This paper intends to fill this gap by examining the effects of leaders on the dynamics of authoritarian regimes. Our investigation focuses on three sets of questions, covering the period from 1946 to 2020. First, do autocratic leaders play a significant role for regime changes, considering the various structural and institutional constraints they encounter? If so, to what extent? Second, in which direction do they matter? Do they matter equally for democratic changes and autocratic deepening? Third, under which political and economic conditions do they exert more substantial influences? The scarcity of large-N studies on this subject is tied to a longstanding methodological hurdle. It is inherently difficult to identify leader traits that influence regime outcomes, let alone measuring them in a reliable way. Opinion surveys that could reveal public values are absent for national leaders. Acquiring behavioral records of leaders is particularly difficult in autocracies due to their nature of secrecy (Barros, 2016).

In this study, we employ an innovative method known as Randomization Inference (RI) for leader effects, developed by Berry and Fowler (2021), to address our research questions. This methodology involves comparing the R-squared statistic derived from regressions using real leader data with those obtained from randomly permuted leader data within the same country. By evaluating the probability of deviation between the actual R-squared values and those generated through repeated simulations, it extracts the leader effects. This methodology circumvents the direct measurement of leader characteristics, assessing leader effects while remaining agnostic about specific leader traits. Compared to conventional regression methods, randomization inference offers several key advantages such as maintaining a decent sample size, reducing serial correlation, and mitigating the mixing of leader effects with transition effects.

Replying on cross-national leadership dataset that covers 150 countries over a span of 75 years, our study yields several major findings. First, the autocratic leaders exert a very substantial influence on regime changes. Democratic leaders, however, do not have as pronounced an impact on regime dynamics, suggesting that democracies bind their leaders more effectively. Second, the autocrats drive regime changes in both democratic and authoritarian directions, indicating considerable individual variation among the “big brothers.” Third, various conditions can amplify or diminish leader effects in autocracies. Factors such as poverty, inequality, the Cold War, and a softer form of authoritarianism are associated with heightened leader effects.

Our study contributes to the understanding of authoritarian regimes in several ways. First, we “bring the leaders back in” to comprehend authoritarian regime dynamics. Existing literature on regime changes has largely focused on structural, institutional, and cultural mechanisms, leaving the role of political leaders understudied. By filling this perspective deficit, we bridge the enduring gap between conditionist and agency perspectives in understanding regime changes. Second, our findings enrich the substantive understanding of the role of leaders in regime changes. By broadening our perspective from leader-death triggered turnovers to systematic leader effects, we arrive at a different conclusion. Our discovery that leaders facilitate both democratic and authoritarian changes also highlight leaders as a genuine variable rather than a mere reflection of objective political interests. Meanwhile, we have found heterogenous leader effects contingent on varying conditions. This differentiation also modifies views from some narrative-based case studies, wherein autocrats seem capable of single-handedly engineering regime changes.

Authors

• Meng Zhu, Tsinghua University

• Wenquan Wu, Tsinghua University

• Kaiping Zhang, Tsinghua University