Referendums, citizens' initiatives, and the quality of public goods: Theory and evidence form Swiss Cantons

What makes governments more responsive and how can we create incentives for them to improve the quality of the public good provided by them? This paper tries to give theoretical and empirical insights into this question, that became salient issues as the role of the qualtiy of governance has been recognised, by particulary looking at what the role of direct democratic institutions could play. We present a model with three parties that are elected via proportional representation. Parties need to form coalitions in order to be able to implement policy. Citizens endogenously decide whether to launch a referendum or a citizens' initiatives. By looking at the cost of this process to the citizens we show that when the direct democratic instituions are more open the legislator may increase his effort to provide the public good. We also find that as the cost goes to zero the medain voter preferred outcome will always be implemented. We test this results empirically by looking at the experience of Swiss Cantons that used such institutions extensively. By looking at infant mortality rates and an index of fatal traffic accidents, proxying the quality of the health sector and infrastructure, we find some empirical support that, after controlling for other factors more openness leads to better public goods. The role of religious and linguistic fractionalization is dicussed, too.