Winners Defend Status Quo; Losers Favor Referendums—Polarization Widens the Gap

Why Voters Choose Referendums?

Bjarn Eck and Emilien Paulis ask whether electoral winners and losers differ in their support for referendums and how affective polarization shapes those differences. The question matters because referendums are a central form of direct democracy: they can be a corrective for dissatisfied voters but also a check on governing parties’ authority.

What the Authors Measure

The study compares attitudes toward referendums among supporters of governing parties (winners) and opposition parties (losers). It also examines affective polarization—partisans' negative feelings toward supporters of other parties—as a potential amplifier of any winner–loser gap.

Cross-National Survey Evidence

• Data come from cross-national public-opinion surveys covering thirteen European democracies.
• Key variables are respondents’ partisan position relative to the government (winner vs. loser), their level of affective polarization, and their expressed support for holding referendums.

Main Findings

• Electoral losers consistently show greater support for referendums than electoral winners across the countries studied.
• Higher affective polarization widens the winner–loser gap in support for referendums.
• Crucially, the polarization effect is driven not by increased demand for referendums among polarized losers but by a decline in referendum support among highly polarized winners.

What This Implies for Democratic Politics

These results challenge a simple narrative that disgruntled losers alone drive demand for direct democracy. Instead, electoral victory combined with partisan hostility can reduce winners’ openness to referendums, with potential consequences for accountability and checks on governing parties. Eck and Paulis’s findings shift attention toward the democratic implications of victory as well as loss, and suggest that affective polarization can undermine citizens’ shared commitment to institutional checks on power.