How Referendums Can Contribute to Democracy, and How They Cannot: Avoiding Unreasonable Expectations

Instruments of direct democracy such as referendums, have always been controversial, including in deliberative democratic theory. It is easy to see why. Referendums are said to emphasize votes over voice, thus encouraging conflict rather than cooperation and providing little room for citizens to actually deliberate with one another. Compared with more recent democratic innovations such as the deliberative poll or citizen’s assemblies, the referendum might then even seem obsolete. These criticisms seem fair enough. Referendum campaigns are not the right context to expect robust face-to-face citizen deliberation, nor do they provide room for compromise or the cooperative search for consensus that deliberative democrats value. Yet, while these criticisms might be right in arguing what referendums cannot do for democracy, I will ask how referendums can still contribute to democratic legitimacy. This more realistic approach, I argue, helps to avoid unreasonable expectations. To do so, I will engage critically with the deliberative ideal and its ambiguous relationship towards voting. On the one hand, the deliberative democratic ideal developed in reference to the concept of the “ideal speech situation”, a counterfactual ideal that can never be fully realised in practice so that the quest for legitimacy is never settled. This leads to a conception of democratic legitimacy which sees forms of political authority, such as voting, as inherently unjustified and stresses instead the importance of keeping the democratic debate open and creating opportunities for citizens to contest authority. On the other hand, deliberative democracy accepts – in Jane Mansbridge’s words - the “importance of getting things done”. In the end, decisions need to be made and solving collective action problems will inevitably require the use of coercion. While this use of coercion can never be fully legitimate, it can still be more or less legitimate. This less demanding approach to deliberative democracy can for example be seen in Bob Goodin’s recommendation of “first talk then vote”. I suggest that the role referendums can play in a deliberative democracy needs to be thought about in reference to this tension between consent and contestation. Referendums do not generate consensus, nor should they. Referendums come into play precisely when the search for consensus has come to an end. Their role then is to act as a fair way of dealing with persistent – but legitimate – disagreement. This has important consequences for how we perceive the legitimacy of a referendum result. Most importantly, the outcome of a referendum should not be seen as reflecting the “will of the people” in any substantive sense. This all too easily leads to the de-legitimisation of dissent. Rather, the referendum should be seen as legitimate for as long as it treats every citizen as individually and morally equal. In order for the referendum to actually live up to this promise, a number of conditions will have to be in place. Most importantly that the referendum campaign itself is sufficiently inclusive and respectful of disagreement. Questions of institutional design have to be geared towards realising these goals.