In this article the ability of Laver and Shepsle's portfolio allocation (PA) theory to account for the survival records of coalition governments in twelve West European parliamentary democracies is examined. The initial test consists of determining whether governments that are in PA equilibrium survive longer than other governments. Simulation experiments are then used to estimate the robustness of these equilibria in order to assess whether robust equilibria have significant survival-enhancing properties. The foil in this investigation is the ideological diversity approach, which sees ministerial accommodation, not ministerial autonomy, as the key to decision making in coalition governments. The evidence favours the latter perspective.