According to theories on coalition-formation, parties with a key position in the coalition game receive higher office and policy payoffs than their coalition partners. In this article, I use two models of government formation – the portfolio allocation model and the political heart model – to identify key players in the coalition game. Both models are modified to incorporate institutional and political constraints on coalition formation, and the predictions of key parties from the four models are compared with the governments that actually formed in five European countries: Austria (1983–2002), Belgium (1985–2003), Germany (1980–2005), Ireland (1982–2002) and The Netherlands (1977–2003). I argue that the modified models are preferred to the original ones on both theoretical and empirical grounds. Using the key parties identified by the modified models, I find that if a coalition member is a key party, then it is able to assert its policy views more effectively than its coalition partners can, but, contrary to expectations in the literature, that it is unable to capture a surpassing share of cabinet offices.