PhD Defense C.C.L. Wagenaar MSc

Referendums most commonly occur in a binary format in which citizens either accept or reject a policy proposal. In practice, many voters are unlikely to be fully satisfied with either of these two options. They may have reservations about the policy or may only be willing to accept it conditional on some amendments. A binary referendum does not allow voters to express preferences for variations on the policy proposal. This may cause voters to reject a proposal despite being in favour of new policy on the topic. Such a rejection also makes it difficult for policymakers to adequately interpret a referendum outcome.

Multi-option referendums could be an effective alternative, as they offer voters a choice between three or more policy options. They result in a single winning option which details the most desired policy route. Compared to binary referendums, multi-option formats have several advantages but also face several challenges, neither of which have been structurally analysed in prior research. Through four academic articles, this thesis explores various design questions relevant to multi-option referendums. The findings contribute to our insights into the implications of design variations for the empowerment of citizens and the materialisation of clear voting outcomes.

This research considers two different phases of the referendum process. First, the agenda-setting phase raises the question of who decides when a multi-option referendum is held and which topics and policy proposals are on the ballot. Second, the balloting phase raises questions on how votes are expressed on the different options and how they are combined into a final result.

The practical manifestations of various advantages and challenges in the two phases are explored by analysing both empirical experiences and new survey data. The empirical research presents a new dataset of national-level multi-option referendums. It discusses what we can learn from how multi-option referendums have been designed and conducted in practice. It also maps different processes of agenda-setting and explores the role of citizens in this phase. The survey research structurally compares the effects of different balloting variations in two separate studies. The first study compares voting on referendum ballots using a single question or multiple questions. The second study compares voting results between a multi-option ballot and a binary ballot.

The research findings imply that multi-option referendums can be an effective instrument for citizen participation when more than two policy scenarios are possible and enjoy support. They can empower citizens both as voters, being offered more choice, and as agenda-setters, enabling them to contribute to the ballot content. Variations in terms of who selects the ballot options and which methods are used to vote on the options each have their own advantages and limitations. The optimal choice may depend on the context and desirable characteristics of the referendum. One general and pressing recommendation for balloting is to use voting methods in which voters can vote for more than one proposal. With such alternative voting methods, voters can either approve of several proposals or rank them according to their relative preferences. This greatly increases the likelihood of yielding a clear majority outcome.

The thesis concludes that though not all referendums may benefit from a multi-option format, the multi-option format should be seriously considered. It maintains the advantages of referendums as an accessible instrument to participate in policymaking on a concrete topic, whilst providing more choice to voters and greater clarity on preferred policy routes. The findings on the characteristics of various ballot designs, voting methods and agenda-setting models can help practitioners and academics to evaluate the relative advantages and limitations of various multi-option designs and to weigh them against the simpler but also more limitative binary referendum design.