Democratic peace arguments make compelling claims about the importance of r
egime type for explaining conflict but do not explain the variation in conf
lict propensity among states that share a common regime type. The authors d
evelop a veto-player approach to capture the effects of executive constrain
t and argue that in parliamentary democracies, cabinet structure and allies
influence states' conflict behavior. A hazard analysis is used to examine
the time before a government's first initiation of force in 569 cabinets in
18 countries from the end of World War I to 1990 (N= 25,238 cabinet months
). Results show that minority governments have a smaller hazard for initiat
ing a conflict than either coalition or majority cabinets. However, the lik
elihood of conflict initiation for coalition and majority cabinets is the s
ame. States with allies are less likely to initiate force. Across states of
relatively uniform political culture, institutional measures of constraint
effectively predict variation in conflict behavior.