In the past, national governments that had negotiated EU membership with Brussels held referendums to legitimate their choice through a popular referendum. Since 2005 national referendums have undergone a paradigm shift: they are not about EU policies. Moreover, the outcome is different. Instead of confirming the decision of their government the majority have rejected the position of the European Union that their national government has endorsed. EU specialists tend to see this as a reaction against policies and institutions of the EU. By contrast, comparative politics specialists are inclined to interpret this as part of an undelrying loss of trust in governments at the national level that can spill up to the EU level. While only a minority of member states have held a referendum, established Opposition parties and protest parties that have never been in government are voicing demands for a referendum on policy issues where the government bound by its participation in the European Council, is defending policies that are unpopular with its own elecctorate, for example, migration and austerity. These demands do not need to produce a referendum vote to affect national and EU policy: Germany is an example of this, while the UK is an example of acceding to the demand on the assumption an absolute majority will trust the government and endorse the EU position. This paper will present a model explaining why the demand for referendums is increasing, and analyse multinational survey data of public opinion in order to test the extent to which the demand comes from people who are support populist views in the pejorative sense or from people who are democrats in Robert Dahl's sense, favouring increased participation in politics. .
