(No) Love Lost? Revisiting Voter-Party Cross-Pressure and Affective Polarization from a Referendum Perspective

This study provides a re-examination of the relationship between voter–party cross-pressure (or incongruence) and affective polarization, defined as the mutual dislike between partisan groups, which poses risks to democratic stability by increasing tolerance for undemocratic behavior, eroding trust, and deepening political divides. While existing research highlights elite polarization as a key driver of affective polarization, recent studies suggest that cross-pressure may instead reduce it by reducing in-party affect and softening out-party hostility. However, these findings often rely on self-reported measures of incongruence, raising concerns about endogeneity and reverse causality, where pre-existing negative attitudes toward one’s preferred party may inflate perceived disagreement. To address these limitations, this paper proposes leveraging referendums as natural experiments to more cleanly identify the causal effect of cross-pressure. In referendum contexts, voters make binary choices on salient issues with clear party positions, allowing vote choice to serve as a behavioral indicator of voter–party incongruence without relying on subjective perceptions. This reduces bias stemming from prior partisan affect and avoids the methodological complexities of instrumental variables or voter advice applications. The empirical analysis focuses on the 2023 Australian “Voice to Parliament” referendum, a highly salient and polarized political event characterized by clear partisan alignments mirroring traditional political divisions and a hostile campaign environment. Using post-referendum survey data from the Australian Constitutional Referendum Study (ACRS), the study focuses on vertical affective polarization, capturing differences in attitudes toward in-parties and out-parties, while excluding horizontal and issue-based forms due to data constraints. The analysis employs ordinary least squares regression to assess in-party affect and linear mixed-effects models to account for repeated observations in out-party evaluations. The study contributes to the literature by offering both a methodological innovation and substantive insights. It provides a cleaner causal framework for assessing the impact of cross-pressure on affective polarization and extends research on referendums, which remain underexplored in this context despite their inherently divisive nature. By focusing on the mechanisms through which referendum dynamics influence partisan attitudes, the paper aims to clarify how issue-based disagreement within party lines shapes broader patterns of political polarization.