Polarization grows alongside increasing uncertainty

Why are political parties drifting further and further apart? And why doesn’t increasing polarization automatically lead to a greater voter turnout? Economist Christoph Kuzmics examines the interaction between political stances, voter turnout, and uncertainty about an election's outcome. With the help of a model of U.S. presidential elections, he aims to explain how these factors influence one another – and what lessons this can teach us about the dynamics of political polarization.
Culture and Society
by Ingrid Ladner
October 22, 2026
Graphic depiction of U.S. ballot boxes on tables labeled “Vote” against a blue-and-red background.
The whole world is holding its breath before the 2026 midterms in the US: How will the elections turn out? An Austrian research team will take a close look at the results.

According to a commonplace theory, political polarization increases voter turnout. This is supported by several examples, such as most recently in Germany, where high turnout rates were reached in Saxony-Anhalt (77.8 percent), and in Hungary (79 percent). With growing political differences between parties, the stakes for the voters rise as well. On the other hand, this does not always necessarily lead to more people going to the polls. In the 2024 U.S. presidential election, voter turnout was lower than before, at 64.5 percent.

Hence, is political polarization actually a reliable driver of higher voter turnout – or is there a more complicated relationship between the two? Economist Christoph Kuzmics of the University of Graz is addressing this question in a project funded by the Austrian Science Fund FWF.

Kuzmics and his team use a new model in order to better understand two classic questions in political economy: how can we explain both the observed level of polarization and the level of voter turnout? For this purpose, the researchers are developing game-theory models of political competition using the U.S. presidential elections as an example.

Polarization and Mobilization

Economist Christoph Kuzmics is developing a completely new game-theoretic model based on the U.S. presidential elections. The model is intended to provide new insights into how polarization, voter turnout, and uncertainty about the election outcome are interrelated. Through this work, the researcher also hopes to gain new insights into voter behavior.

A researcher wearing a blue T-shirt is standing in front of a research institute building.
Microeconomist Christoph Kuzmics wants to better understand why parties are increasingly drifting apart. © Jörn Kleinert

Mr. Kuzmics, why did you choose the U.S. electoral system?

Christoph Kuzmics: The U.S. presidential election is of great significance for Europe, as its effects stretch far beyond the United States. At the same time, the political system there lends itself quite well to research: with two dominant parties, political stances are easier to categorize than in Europe, where many parties and coalition negotiations produce a more complex political landscape. This means that certain things can be observed much more clearly in the USA.

You’re developing a simplified model of U.S. presidential elections based on three factors: political polarization, voter turnout, and uncertainty about the election outcome. What can such a model reveal that couldn’t be discerned from past election data?

Kuzmics: The unique aspect of our approach, which is also a first, is our focus on the interplay of all three factors together, since they are mutually influential. That is to say, we look at the mutual impact between the positions parties take and also at who ends up voting – that’s something we already understand to some extent. But the big unknown factor is the uncertainty about the election outcome. It affects both whether people go to the polls and what position parties adopt.

What assumptions about the behavior of parties and voters are important for your model?

Kuzmics: In our model, parties take a position on a political spectrum between left and right. Voters naturally choose the party whose positions are closest to their own. But the central question is why do people go to the polls, and why do some stay home? Accordingly, parties need to consider not only their own position, but also how to mobilize their supporters to actually vote. Our model examines what incentives play a role in this process.

You’re still in the early stages of the project – what are your hypotheses so far?

Kuzmics: Older models assume that parties mainly compete for the votes of centrist voters. Consequently, it would make sense for the two major U.S. parties to position themselves in the political center to reach as many voters as possible. Instead, they are drifting further and further apart. However, the old models do not take sufficient account of the fact that voters do not automatically go to the polls, which is why greater polarization would not necessarily lead to higher voter turnout. In the U.S. in particular, this correlation cannot be clearly observed. Our model aims to explain why that is the case.

“The question of why people vote is still puzzling for researchers.” Christoph Kuzmics

A central aspect of the project is uncertainty about the election outcome. Can you explain how a change in uncertainty affects voter turnout and the parties’ policy stances?

Kuzmics: The question of why people vote is still puzzling for researchers. A weighty factor is the belief that one’s vote can make a difference. But many voters do not believe that their ballot will influence the outcome. This is especially true when the outcome seems fairly certain. The more uncertain the election outcome, the stronger the potential incentive to vote. Another mobilizing factor may arise when opinion polls suggest a clear outcome but the actual odds are difficult to assess. When Donald Trump first ran for president in 2016, the opinion polls saw him as the clear loser, and yet betting odds suggested he stood a chance.

How do you define uncertainty in your model?

Kuzmics: Based on previous studies, we assume two effects that we want to examine: first, greater uncertainty about the election outcome tends to lead to lower voter turnout. This indirectly triggers the effect of polarization. When general uncertainty is high, the parties take more clearly distinct stances, which motivates more people – including those on the fringes of the political spectrum – to go to the polls. We want to find out which of these effects is predominant.

What influence can opinion polls have on uncertainty, and how do the parties react to them?

Kuzmics: Opinion polls provide a fairly good indication of which party people support. As I’ve said before, the more difficult thing is to predict who will actually vote. Parties can respond in two ways: on the one hand, they can shift their political positions to appeal to new voter groups. We’re seeing this now ahead of the midterms, where some Republican candidates are distancing themselves from Trump and running on a more moderate platform. Alternatively, they can try to mobilize their existing base more effectively. We’re investigating which of these two strategies carries more weight.

“Campaign messages will not do much if the political platform or economic track record isn’t convincing.” Christoph Kuzmics

American election campaigns receive vast financial donations designed to provide incentives for voters. Does the money factor also come in here?

Kuzmics: Initial results from our model suggest that the parties’ policy stances are more important than the mobilization of their supporters. The state of the economy also plays a central role: campaign messages and high spending on advertising will not do much if the political platform or economic track record isn’t convincing.

Is there a point at which political polarization no longer mobilizes additional voters?

Kuzmics: The model shows this relatively clearly: political polarization has its limits. Parties must win majorities. If they move too far to the fringes, they make it easier for the competitor to appeal to voters in the center. In addition, greater polarization among politicians does not necessarily mean that the population is more divided as well. Highly polarized elections are fundamentally problematic because they are extremely inefficient: policy measures end up being significantly divorced from what the majority of voters want, regardless of who wins.

In the second part of the project, you will use publicly available data on past U.S. presidential elections to assess how well your model captures reality. How much simplification does the model require?

Our flexible model assumes that voters are guided by what matters to their respective political camp. At the same time, individual factors play a role, such as the effort involved in going to vote. While the model simplifies reality, it is nevertheless surprisingly good at explaining voter turnout in various U.S. states: the calculated values match the observed data by about 80 percent. This shows that the approach is helpful, even if we still need to better understand the personal motivations of voters.

How can your research contribute to our understanding of voting behavior in other countries?

Kuzmics: In the U.S., one of the two major parties is largely free to implement its political agenda on its own after winning an election. In Europe, on the other hand, parties usually have to form coalitions and make compromises. As a result, the policies actually implemented may turn out to be more moderate, even if individual parties hold strongly polarizing positions. Our model can facilitate a better understanding of these differences between political systems.

What insights and possible conclusions do you hope to draw once the project is complete?

Kuzmics: We would like to provide a coherent explanation for the correlations observed in the past between voter turnout, polarization, and election results in U.S. elections across U.S. states and over time. This will help us to better understand why parties are increasingly drifting apart. Ultimately, it will also allow us to identify potential avenues for policy recommendations. If, for instance, uncertainty is closely linked to polarization, then our model can be used to analyze the effects of regulatory measures – such as those related to voter turnout or targeted advertising on social media – which may be a potential cause of electoral uncertainty.

About the researcher

Christoph Kuzmics is a professor of microeconomics at the Institute of Economics at the University of Graz. He previously held positions at Bielefeld University and the Kellogg School of Management at Northwestern University in Illinois. His research focuses on decision theory, game theory, mechanism design, and the theory of political processes. The project “Polarization and Mobilization under Uncertainty” is funded by the Austrian Science Fund (FWF). It is set to run through 2029.