Why does it take some leaders more time than others to dismantle democracy? In some cases, such as Turkey, it took more than a decade before democracy broke down, whereas in other cases as in Hungary, the process lasted only a few years. The literature on democratic backsliding is full of insights on how leaders elected in free and fair elections undermine democracy from within to remain in power. However, current studies so far have overlooked whether and how incumbents’ political parties matter for democratic backsliding onset and subsequent democratic breakdown. There is an implicit assumption that incumbents’ parties facilitate backsliding or that they do not matter altogether in the case of populist or personalist leaders (Weyland 2024; Rhodes-Purdy and Madrid 2020). The paper bridges this notable gap by making two central arguments. First, incumbents’ political parties facilitate backsliding onset but increase the duration of backsliding due to increased uncertainty and factional politics. Second, the paper argues that democratic breakdown after backsliding hinges on the ability of incumbents to de-institutionalize their parties. The paper examines seven hypotheses on the relationship between the incumbent’s partisan identity, party institutionalization, democratic backsliding and subsequent breakdown. We employ time-variant, multivariate survival regression analysis for autocratization episodes in all electoral and liberal democracies between 1945 and 2020 using data from V-Dem, the Ideologues dataset among others. The results confirm the importance of the incumbent’s party and its de-institutionalization for the onset and duration of backsliding and the subsequent breakdown of democracy.