Leadership transitions and survival: Coups, autocoups, and power dynamics

This thesis addresses a significant gap in the literature on irregular leadership transitions by systematically integrating autocoups—cases where incumbent leaders extend their constitutionally mandated terms through extra-constitutional means. It refines the conceptual definition of autocoups, resolving existing ambiguities to align them more closely with conventional coup frameworks. Based on this refined definition, the thesis presents a novel global dataset of autocoup events from 1945 to 2023, encompassing 83 documented cases, of which 64 were successful. Using this dataset, the study conducts a large-N analysis to identify the structural determinants of autocoups. The findings reveal that regimes with concentrated executive power, such as presidential democracies and personalist regimes, are significantly more prone to employing autocoups as a strategy for power retention compared to other regime types. This pattern contrasts with traditional coups, which have historically been more prevalent in military regimes. The analysis subsequently examines leadership survival, utilising survival analysis techniques to evaluate the political longevity of leaders who assumed office through regular means, traditional coups, or autocoups. Contrary to the hypothesis that the mode of power acquisition significantly influences leadership survival—particularly the expectation that autocoup leaders would enjoy longer tenures than those installed through traditional coups—the results suggest that, when excluding very short-lived tenures (i.e., less than 180 days), the mode of accession does not exert a statistically significant effect on leadership duration; instead, regime type emerges as the primary determinant: leaders in military and personalist regimes exhibit significantly higher hazard ratios for removal compared to the reference category of dominant-party regimes, consistent with trends observed following traditional coups. The thesis further explores the broader institutional consequences of irregular power transitions, focusing on their impact on democratisation. Using Polity V scores as a proxy for democratic quality and applying a country-fixed effects model, the analysis finds that autocoups are associated with a sustained decline in democratic indicators, both immediately after the event and over a two-year period; in contrast, traditional coups exhibit a "U-shaped" effect: although Polity V scores decline sharply in the immediate aftermath, they typically recover to pre-coup levels within two years. These findings highlight the divergent political trajectories triggered by coups and autocoups, underscoring the need for increased scholarly and policy attention to the persistently detrimental impact of autocoups on democratic governance. Collectively, these findings illuminate the distinct characteristics, causes, and impacts of coups and autocoups. The research makes several substantive contributions: it clarifies the conceptual boundaries of autocoups, provides a new empirical foundation for their systematic study, and offers robust comparative insights into how different modes of irregular power transition affect leadership survival and institutional development. These findings have significant implications for academic scholarship and policy-making, particularly in the context of global democratic backsliding and the resilience of political institutions.