On the whole, there is a sharp disconnect between theoretical accounts of the government formation process and the empirical models that seek to test it. Existing studies inappropriately examine the choice of PM party and the choice of governing coalition in complete isolation from one another. In fact, the government formation process is sequential and nested in nature -- first a formateur is determined, and then the formateur goes on to try to form a government that contains her own party. The choice of formateur is endogenous, both affecting and affected by potential government coalitions. In this paper we apply a new method -- a mixed nested logit -- to the study of the government formation process designed to account for this process, narrowing the gap between theory and empirics, and allowing us to test several hypotheses that cannot be tested under previous empirical approaches, such as whether potential governments are more likely to form if they are ideologically closer to the formateur. We evaluate our hypotheses using a new data set that we constructed containing information on almost 190,000 potential governments drawn from 402 government formation opportunities in 17 parliamentary democracies from 1945 to 1998.
