Regional government formations and the benefits of coordination

One of the main features of democracy is that citizens can vote the incumbents out of office. This accountability mechanism should guarantee that governments are responsive to the preferences of its citizens. However, in parliamentary systems, voters do not elect governments, but voters elect representatives who appoint a government. Whether government formations are responsive to election results is thus not a given. Especially on the regional level this link might be obfuscated because parties have incentives to form governments that are congruent with the national government (i.e., regional governments have the same composition as the national government), especially when regions can co-determine national policies (Bäck et al., 2013). Increasing regional involvement in national policy making might thus come at the cost of how responsive regional government formations are to regional election results. In this paper we analyse the likelihood that regional governments will be formed and how responsive these formed governments are to the election result. We expect that in regions that exercise high levels of shared rule together with the national government, governments are more likely to be formed if they are congruent with the national government. As parties can benefit from vertical coordination. Meanwhile, if a region co-determines policies together with other regional governments, regional government are more likely to be formed if they are congruent with regional governments that are formed in other regions. As parties here can benefit from horizontal coordination. Finally, if the benefits from coordination are small (low levels of shared rule) government formations are more responsive to the election results. To test these expectations, we utilise a dataset that includes 3.926.000 potential coalitions and a total of 2.212 formed governments in two hundred regions in 13 Western European countries where the regional government is dependent on the support of the regional parliament. First, we find that the governments that are formed on the regional level are not that different from those at the national level (i.e., governments are more likely to be formed if the coalitions are minimal winning and ideologically congruent). Furthermore, we find that the extent to which parties can benefit from either horizontal or vertical coordination affects which government will be formed. If the benefits of coordination are limited, government formations become more responsive to the regional election result – to the extent institutional rules allow this.