How representative are referendums? Evidence from 20 years of Swiss referendums

Direct democracy allows citizens to reverse decisions made by legislatures and even initiate new laws which parliaments are unwilling to pass, thereby, as its proponents argue, leading to more representative policies than would have obtained under a purely representative democracy. Yet, turnout in referendums is usually lower than in parliamentary elections and tends to be skewed towards citizens of high socio-economic status. Consequently, critics of direct democracy argue that referendum outcomes may not be representative of the preferences of the population at large. We test this assertion using a compilation of post-referendum surveys encompassing 148 national referendums held in Switzerland between 1981 and 1999. Uniquely, these surveys also asked non-voters about their opinion on the referendum's subject. Comparing opinion majorities in the surveys against actual referendum outcomes we show that representativeness increases slightly in turnout as well as over time. However, we find only few cases where the outcome would have been more representative even under full turnout vis-a vis a counterfactual representative outcome. Thus, our results are in line with research on the turnout effect in elections: Higher turnout would not radically change the outcome of votes. On balance we find more cases where referendums provided more representative outcomes than cases where the outcome was unrepresentative vis-a-vis representative democracy. Hence, we conclude that, overall, direct democracy seems to improve representation in Switzerland.