The International Diffusion of Democratic Backsliding

This paper examines the diffusion of processes of democratic backsliding across national borders. Studies of democratic backsliding focus overwhelmingly on domestic-level explanations. However, two observations imply an important international dimension of democratic backsliding: First, backsliding countries form clusters. Second, leaders of backsliding countries seek to cooperate with leaders from other backsliding countries. Scholars studying international diffusion of political regimes have highlighted the diffusion of both, democracy and autocracy, but have yet to consider democratic backsliding. I argue that democratic backsliding diffuses across borders. Each case of democratic backsliding increases the risk of backsliding in future cases by generating an air of appropriateness and feasibility, and by setting in motion mechanisms of emulation and learning on the part of political actors. The process of diffusion of democratic backsliding is distinct from the diffusion of autocracy, in that it is the precedent of the process of backsliding – rather than the precedent of autocracy abroad – that induces diffusion. I test the proposition using spatial lag models and an original measure of democratic backsliding and the diffusion of backsliding, based on data from the Varieties of Democracy project. Spatial lag models are apt to capture the complex dynamics of diffusion across time and space and are frequently employed to model diffusion processes.