Government Formation and Veto Points: Generalising the Lijphart-Sjölin Conjecture

The Lijphart-Sjölin conjecture states that strong legislative veto points increase the size of coalitions. Existing studies have either rejected the conjecture (Volden and Carruba 2004, AJPS), or found corroborating evidence based on simple cross-section regressions (Ganghof 2010, BJPS) or solely driven by second chambers without weak veto power (Druckman, Martin and Thies [DMT] 2005, LSQ). Building on the latter we argue that a generalized version of the conjecture is corroborated by the data. Our study departs from DMT in four main ways. First, we greatly extend the sample size to 21 advanced democracies and include more countries with powerful veto points such as Australia and Japan. Our dataset comprises 560 government formations between 1945 and 2010. Second, we include veto points other than second chambers such unicameral supermajority rules in Finland (until 1992). Third, we test our hypothesis using mixed logit models. Fourth, we use the results of the empirical model to visualize the effect of veto points on the probability of forming oversized coalitions.