The charm of independent voters
2025
Read The charm of independent voters (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2512-17085.txt.
Introduction
Ernő Buzás and Attila Szilva model the charm of independent voters, a flexible minority that can sway polarized elections. Using a modified Deffuant model, where agents shift opinions toward those they meet within a tolerance, they vary how open agents are to switching.
Even simple splits in openness spawn rich outcomes. Open-minded agents can block full convergence into party blocs, give rise to short-lived centrist clusters, or line up with major parties in the end. Small shifts in openness values reshape the whole map of political competition. The patterns mirror real independent voter behavior and offer a simple account of swinging outcomes and unstable centrist coalitions.
Catalog record
| YEAR | 2025 |
|---|---|
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 21 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2512-17085
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
The charm of independent voters Abstract Independent voters play an increasingly decisive role in contemporary elections, yet their collective behavior remains poorly understood. This paper investigates how a minority of voters with greater flexibility in their political preferences influences opinion formation in polarized electorates. Using a modified Deffuant model, we show that even simple heterogeneity in agents’ openness to vote switching can generate rich and unexpected collective outcomes: "open-minded" agents may (i) prevent full convergence into established party blocs, (ii) give rise to transient centrist clusters, or (iii) align with the positions of major parties. These dynamics resemble empirical patterns observed among real-world independent voters. Our results demonstrate that small shifts in openness parameters can substantially reshape the macroscopic structure of political competition, offering a simple explanation for oscillatory electoral outcomes and the emergence of unstable centrist or cross-cutting coalitions. Ernő Buzás, Attila Szilva 1 Introduction In the past few years a number of Western countries experienced an unusual level of change in political life, which might extend into the next few years. Established parties lost ground and new parties and movements experienced rapid growth. Examples include the rise of Donald Trump and the MAGA movement in the United States, Giorgia Meloni in Italy or the 2024 presidential election shock in Romania [ 25 , 14 , 16 , 1 ] . We see events that are hard to interpret for pollsters and political scientists [ 21 , 4 , 18 ] . At the same time, it has been found that polls asking respondents about the views of their closest peers can improve prediction accuracy [ 10 ] . Moreover, so-called opinion dynamics (OD) models [ 23 , 27 , 26 , 12 ] can accurately predict individual opinions in online populations, as shown in Ref. [ 30 ] . The emergence of these new phenomena may reflect that a growing share of the electorate has become more willing to switch their votes. The specific reasons for this willingness can be numerous and quite different, from economic changes in society to demographic, cultural, religious shifts. Corruption scandals and the emergence of new (social) media platforms [ 20 ] can also facilitate the rapid rise of new political actors in the vacuums left by the fall of the previous leaders. Although the overall picture is complex, we focus solely on modeling voters’ willingness to change their opinions through interactions with others, which we assume to be a key underlying mechanism [ 6 ] . While identity-based societal shifts, such as cultural or religious ones are likely to cause systematically biased changes in opinions (such as leaning towards one political side), the behaviour of independent voters whose behavior resembles the Big Five trait of Openness [ 22 ] , could be really relevant to study the collective effect of "flexible" individuals. Of the Big Five traits, openness and conscientiousness are the most likely to increase the likelihood of vote switching [ 13 ] . Another relevant category concerns a much larger share of the electorate. These are those who identify as "independents", especially in two-party systems such as the United States of America, where their recently growing share might have become as high as 40-46%. They do not conform to how partisan voters behave and show significant volatility in their voting preferences over time [ 24 ] . Since a real life independent voter would not stick to a given party or candidate over time, our model captures this behaviour as well. While there can be many complex real-life reasons why individual voters depart from established parties and become more susceptible, or more "open-minded", toward extreme or non-mainstream views, we will control the open-mindedness with a single parameter in our computational model, which will be an extension of the Deffuant model [ 6 ] , a seminal OD model. Note that the adoption of OD as a modeling framework is motivated by empirical findings highlighting the importance of social contagion in political opinions, as reported in Ref. [ 10 ] . The paper is organized as follows. In Section 2 we present a theoretical overview of OD models, starting from the classical voter model. In Section 3 we introduce the basic Deffuant model with random but uniform starting opinion distribution, while in Section 4 the initial distribution of opinion is non-uniform. In Section 5 we add a few agents with higher open-mindedness and study how they change the overall outcome. In Sections 6 and 7 , we discuss the meaning of these results and conclude by outlining their possible real-world applications. 2 Theoretical background To understand voter behavior, we developed a computational model in the theoretical framework of opinion dynamics. OD is an interdisciplinary field that studies how individuals’ opinions evolve through social contagion and other interaction processes, revealing how populations polarize, converge toward consensus, or stabilize in a state of coexistence [ 23 , 27 , 26 , 12 ] . Opinions can be one- or multi-dimensional, binary or continuous, depending on both their object and the way of measurement, and OD deals with their multi-state dynamics on complex networks. The dynamical equations in OD are similar to interaction models known in magnetism [ 28 ] and statistical physics [ 3 ] . Historically, the first OD model is the classical voter model (VM) [ 5 , 15 ] in which each individual (voter) can assume one of two discrete states (e.g., yes/no, up/down, Democrat/Republican, o_{i}=\pm 1 ), and a single voter resides at each node of an arbitrary (static) network. A voter at node i is selected at random and copies the state of a randomly chosen neighboring voter ( j ). This dynamics is repeated until consensus is necessarily reached. Another seminal early OD model is the DeGroot model in which the opinions of the agents are represented by continuous real numbers and each agent updates their opinion by forming a weighted average of their neighbors’ current opinions [ 7 ] . While the classical VM is stochastic and handles the diffusion of categorical labels, the DeGroot model is deterministic and describes the diffusion of continuous signals. Nevertheless, the essence of the models is the same: agents update their states by adopting the states of their neighbors. Note that the classical VM is one of the few exactly solvable many-body interaction systems when voters are situated on a regular lattice or a homogeneous graph, and it can be shown that the DeGroot model describes the mean-field evolution of the classical VM [ 12 ] . Some models incorporate mobility due to daily commuting, i.e. it is possible to calibrate them to real socio-economic data [ 8 , 17 ] . There are examples when VMs can accurately predict individual opinions in online populations [ 30 ] . A great summary of the binary-state dynamics on complex networks is given by Ref. [ 11 ] and a mini-review about reality inspired VMs has been published by Ref. [ 23 ] . While VMs capture social influence, they do not incorporate homophily [ 29 ] , the well-documented tendency for individuals to connect with others who share similar characteristics or views. One way to take into account homophily is to