Divide and Conquer: Presidents, Parliaments, and Political Polarization during Electoral Campaigns

Although political polarization is in the center of contemporary debates on democratic backsliding, the institutional determinants of this phenomenon remain understudied. In this paper, we investigate how the form of government can affect polarization levels in a country. Specifically, we focus on the role of executive elections as the primary mechanism. Presenting both a formal model and cross-national empirical analyses, we show that executive elections under presidential regimes are significantly more likely to affect political polarization among citizens. Candidates in parliamentary systems are often party leaders who are committed to party policies and cannot deviate much to disclose new information. In contrast, presidential candidates can strategically design the discourse in their election campaigns, thus influencing voters' opinions and making mass polarization more likely. Our empirical analysis shows that this effect is nonlinear and especially profound when the initial polarization in a country is low.