DP14906 Random Power to Parties and Policies in Coalition Governments
We exploit a natural experiment involving a randomization of votes across parties within coalitions in all local elections in Italy for over a decade. A lottery on the position of party symbols in the ballot papers coupled with a behavioral focal point allows estimating the average treatment effect of parties for coalition policies. A non-marginal boost of votes shifts budgetary spending towards the treated party's platform, but only for issues that are salient in that party's political manifesto. The effects are stronger for more divisive electoral issues like welfare, taxes, and security. The mechanisms relate to perturbations of legislative representation within majorities that spills over to the appointment of cabinet members affiliated with parties. Empowering different parties leads to a selection of policymakers with different socio-demographic characteristics (particularly gender, age, and occupation). The unintentional experiment helps shed new light on the role of parties and their ideologies.
