Do Close Referendum Losses Demobilize Voters? Evidence from Swiss Municipal Voting

Abstract

Does experiencing a local referendum loss affect subsequent voter turnout? I exploit the unique setting of Swiss direct democracy, where municipalities vote on federal referendums that may pass nationally despite local opposition. Using a regression discontinuity design at the 50% vote share threshold across 2,122 municipalities and 56 federal referendums from 2010–2019 (with outcome data extending through 2022), I compare subsequent turnout in municipalities that narrowly "lost" (voted against a passing measure) versus those that narrowly "won" (voted for a passing measure). I find no evidence that local referendum losses affect subsequent turnout: the RDD estimate is 0.05 percentage points (SE = 0.84, p = 0.95). The McCrary density test shows no evidence of manipulation (p = 0.92), and the null result is robust across bandwidth choices, polynomial specifications, and placebo cutoffs. This finding suggests that Swiss direct democracy is resilient to "sore loser" demobilization effects—voters continue participating regardless of past local outcomes. The result contributes to the policy feedback literature by demonstrating that repeated referendum experience, even when locally unsuccessful, does not erode democratic participation.