Effect of Electoral Seat Bias on Political Polarization: A Computational Perspective

TL;DR

This paper tackles the problem of whether electoral seat bias drives political polarization by using an agent-based Monte Carlo simulation in a $2$-dimensional policy space to isolate the effect of large-party bias. It integrates realistic voter behaviors (strategic voting, bandwagon effects, retrospective voting, thermostatic and affective shifts) and employs seat-votes mappings for Jefferson–D'Hondt and FPTP to study polarization outcomes across repeated electoral cycles. The main findings show that stronger large-party bias yields higher long-run polarization, with robust evidence across parameter variations: polarization declines as district magnitude $m$ increases under Jefferson–D'Hondt, and increases with power-law exponent $eta$ under FPTP, both supporting the causal role of seat bias. The work highlights a potentially actionable channel for policy reform, while acknowledging limitations such as the exclusion of party-system change and turnout dynamics, suggesting future work to explore dynamic party evolution and coalition formation. These insights contribute to understanding how institutional design can influence polarization dynamics in democratic systems, offering computational evidence that seat-bias-reducing reforms may attenuate partisan fragmentation.

Abstract

Research on the causes of political polarization points towards multiple drivers of the problem, from social and psychological to economic and technological. However, political institutions stand out, because -- while capable of exacerbating or alleviating polarization -- they can be re-engineered more readily than others. Accordingly, we analyze one class of such institutions -- electoral systems -- investigating whether the large-party seat bias found in many common systems (particularly plurality and Jefferson-D'Hondt) exacerbates polarization. Cross-national empirical data being relatively sparse and heavily confounded, we use computational methods: an agent-based Monte Carlo simulation. We model voter behavior over multiple electoral cycles, building upon the classic spatial model, but incorporating other known voter behavior patterns, such as the bandwagon effect, strategic voting, preference updating, retrospective voting, and the thermostatic effect. We confirm our hypothesis that electoral systems with a stronger large-party bias exhibit significantly higher polarization, as measured by the Mehlhaff index.

Figures (24)

• Figure 1: Running example: initial party sizes and positions. Parties are represented by colored circles centered at ideal points and with area proportional to initial size. Simulation parameters correspond to those given in the next section.
• Figure 2: Running example: first iteration. Seats were allocated using the Jefferson--D'Hondt method with $m = 12$. Colored circles represent parties after the first election (area of each party is proportional to its seat share), while colored dots represent a sample of voters. Note that there are voters without corresponding party circle -- they have voted for a party that failed to win any seats due to the natural threshold.
• Figure 3: Running example: second iteration.
• Figure 4: Running example: last iteration.
• Figure 5: Effective number of winners as a function of district magnitude ($m$). Note that smaller district magnitude implies greater seat bias in favor of larger parties.