Despite the rich literature on coalition politics, we know surprisingly little about how coalition governments actually work: how do government parties reach agreement on policies, and how do they resolve conflicts among them? Designing mechanisms, norms and rules that make coalition governments work is crucial, because coalition governments constantly face threats and external shocks that may lead to an early government termination. In this paper, we study what kind of mechanisms of mutual control government parties employ to enforce joint action. We argue that (1) these mechanisms are unlikely to be effective if used alone and (2) that political parties thus devise complementary and synergetic coalition governance mechanisms. Thus, coalition builders are concerned with designing – what we term – the architecture of coalition governance. Based on theories of delegation and accountability we propose a unified approach and derive several ideal typical configurations of coalition governance mechanisms. Drawing on data from coalition governments in Western and Central Eastern Europe, we use Bayesian Configuration frequency analysis to analyse the prevalent empirical configurations. Here, we find that coalitions indeed choose mechanisms as substitutes and supplements. We then explore the effect of different configurations of governance mechanisms on cabinet duration in Western and Central Eastern Europe. Our findings highlight the importance of factors at parties’ discretion for cabinet duration and provide promising avenues for future research.
