In invoking the confidence procedure, leaders willingly jeopardize their job. To avoid replacement, they need to secure unanimity within their cabinet before calling the vote and the support of the majority of legislators voting on the motion. This begs the following questions: under what circumstances will PMs be more likely to initiate a confidence vote and how do these circumstances affect the prospects of success of the vote? This puzzle is non-trivial because these votes have been traditionally used for salient and controversial matters. In this project, I investigate how institutional features that shape the onerousness of governing affect the PM’s choice to employ this device and the outcome of the vote. There are factors that can make it harder for the PM to obtain consensus from her ministers and parliament. I identify these as the size of the coalition, its degree of heterogeneity, the share of the coalition controlled by the PM, the size of the opposition, and the cabinet’s minority status. By relying on a novel dataset of PMICVs events, I outline different instances of use of confidence votes initiated by PMs and I present an empirical study of the determinants of their usage and success.