On the eve of the interwar period, a set of interrelated structural factors (i.e. socio-economic development and state- and nation-building processes) had created a distinction between resilient democracies installed before the end of World War I and fragile new democracies. Our theoretical framework proposes that these deep structural conditions were linked to interwar democratic survival via a set of more proximate factors, namely, democratic legacies and a strong associational landscape (i.e. vibrant civil societies and institutionalized parties). These factors increased the likelihood that political actors would comply with the democratic rules of the game via a number of mechanisms, such as democratic learning, mutual trust among key political actors, and the capacity to channel the ideas and frustrations of opposing societal groups via the established political institutions. Anti-democratic forces had only little support and room to manoeuvre in cases with firm democratic legacies and associational landscapes. On the contrary, where these factors were absent, democracies broke down after a few years. Finally, where they were partly present, democracies went through prolonged crises only to eventually break down, often during the Great Depression of the 1930s.