Zero-sum Democracy? The consequences of direct democratic victory and loss for political support in comparative perspective

Direct democratic decision-making is still on the rise, and often associated with it is the presumed capacity to generally alleviate perceived legitimacy deficits of representative democracies among both the wider public and participants. Yet, recent research more and more questions this general ability, which was often presumed based on participatory and procedural fairness theories. Rather, it is argued, that outcome favorability, and therefore the division into winners and losers of direct democratic and other participatory procedures, is the primary driver of the support-increasing consequences of more direct citizen involvement. We extend this theoretical perspective for the case of direct democratic procedures within established representative democracies: Building on the game-theoretic differentiation between political zero- and positive-sum games as well as theories on socialization into the role of political procedural losers, we argue, that binary direct democratic procedures in general, and especially within contexts with low levels of direct democratic experience are prone to producing sore procedural losers. Both the polarizing nature of these procedures, as well as the limited experience with the status as a direct democratic loser within a given context, can reduce the incentive and the ability, to display stable levels of support for those who received unfavorable outcomes and lead to a retraction of political support. We test our hypothesis based on 10 individual level panel datasets over the course of regional and national referendums in six different countries, using different standard measures of political support as dependent variables. In a second step, we compare the impact of referendums and their outcomes to that of elections within the same contexts. Results from fixed effects panel regression models for both, the wider public, as well as those participating in the procedures under study, show, that direct democratic procedures are indeed prone to producing sore political losers, with losers among the wider public reducing their previous level of support in half of the cases. While winners largely remain stable or, as it is most often the case, even increase their support, losers only very rarely increase their level of political support after referendums. Additional analysis shows that the resulting winner-loser-gaps in supportive orientations often persist for a long time and are comparable in scale to those created by electoral contests. Our comparative analysis has far-reaching consequences for the often-assumed capacity of participatory processes to address legitimacy deficits within representative democratic systems, as well as our understanding of the impact that these kinds of procedures have on the political attitudes of democratic citizens.