In my paper, I address the problem whether policy position of parties explain coalition composition of Hungarian cabinets formed in 1990-2010. To study the impact of policy on government formation, I test four policy-based hypotheses of „Top 10” issues (Bara 2006), tangential profile (Enyedi–Körösényi 2001), „silent partner”, and „the least evil”. I investigate how well potential coalitions, based on these hypotheses, predict historical coalitions formed during the period between 1990 and 2010. Institutional context of government formation (formateur party, recognition rule, positive parliamentarism) will be applied in hypothesis testing, while policy positions will be calculated from Comparative Manifestos Project data using scaling method of Lowe et al. (2011). Previous research (Horváth 2011) shows that predictive power of policy positions in government formation increases between 1990 and 2002. I expect this tendency to continue in 2002-2010. I try to provide a tentative explanation for this tendency in my paper. In doing so I will consider factors such as mechanical effect of party system concentration (reduced number of potential coalition partners; joint election manifestos), reduced policy distances due to party system polarization, previous experience in governing together as coalition partners.
