������Ƶ

Partisan polarization in the United States now seems to go beyond ideological or policy differences to deep resentment and dislike across party lines. The negative implications of partisan hostility became evident during the divisive 2020 election campaign and the subsequent violent insurrection at the U.S. Capitol. It is no surprise that in President Biden’s inauguration speech he implored Americans to “show respect to one another” and reminded his nation that “politics need not be a raging fire destroying everything in its path.”¹ This essay highlights recent research on how alternative electoral institutions can promote cooperation among political elites and reduce out-party hostility and negative partisanship.

While the bulk of research on out-party hostility and the related phenomenon of affective polarization has been focused on the United States, the emerging comparative scholarship leverages the existence of different institutional arrangements to analyze its structural underpinnings.² A consistent finding is that voters in more proportional electoral systems show less hostility toward out-partisans.³ Figure 2 below demonstrates that relationship by plotting the average level of out-party dislike in a country against its average logged district magnitude—meaning the number of electoral seats assigned to each district within a country.

In a paper co-authored with James Adams and Noam Gidron, we built on previous research to examine how electoral systems influence partisan resentment by focusing on party coalitions. We explored two ideas: first, that citizens feel more positively toward parties that co-govern with their preferred party, regardless of that party’s policy stances; second, that the emotional benefits of co-governance persist, with positive feelings lasting even after the coalition ends.

The first argument, that supporters of parties in coalitions feel more warmth toward their coalition partners, is intuitive. Citizens often believe co-governing parties have more similar ideologies than their manifestos suggest, which can increase positive feelings toward out-parties.⁴ Partisans witness their party and coalition partners defending government performance against opposition and media criticism. This public display of mutual support likely enhances partisans’ perceptions of coalition partners’ character, leading to warmer emotional evaluations. This suggests that politics is less “us versus them” or zero-sum in proportional systems. A recent study by Lotem Bassan-Nygate and Chagai Weiss uses a clever survey experiment which takes advantage of the ambiguity surrounding coalition formation in the 2019 Israeli elections—priming voters to expect party cooperation in the formation of a unity government promoted tolerance across partisan lines, even in the hyper-polarized Israeli setting.⁵ Our own results, based on observational data, suggest that this tolerance promoting effect of coalitions generalizes to western democracies more broadly.

Our second argument is less intuitive. We argue that there are at least three reasons to expect that the impacts of coalitions will be durable even once the coalitions themselves have dissolved.

• First, party identification, as Morris Fiorina describes, can be thought of as a “running tally” of citizens’ evaluations of parties’ policies and performance. This can encompass events spanning decades and may influence partisans’ out-party evaluations long after the coalition ends.⁶
• Second, David Fortunato and Randolph Stevenson find that past co-governance impacts citizens’ current perceptions of party policies, causing them to overestimate policy affinity between their party and former coalition partners. We expect a similar impact on partisan affect.⁷
• Third, past co-governance may affect party elites’ public interactions with former coalition partners. Opposition party elites may maintain cooperative relationships to signal a willingness to co-govern in the future. These signals can impact citizens’ perceptions of their own party and previous coalition partners.

To understand why the lingering impact of coalitions may potentially have a more powerful impact on a country’s level of partisan hostility than solely focusing on a country’s current coalitions, consider that while only 8 percent of current party-pairs in our sample share in governance, another 16 percent of parties have co-governed at some time in the last twenty years.

For a stylized example, consider the United States and the Netherlands. The Democratic Party, in a winner-take-all electoral system, experiences zero-sum politics with little bipartisan cooperation. Conversely, the Dutch Christian Democratic Appeal (CDA), a center-right party, operates in a highly proportional electoral system, leading to frequent multiparty coalitions. For instance, prior to the 2010 election, the CDA had governed in various coalitions with parties spanning the political spectrum, including the center-left Labour Party, the center-right People’s Party for Freedom and Democracy, and the centrist liberal Democrats 66. Unlike the United States, where no two parties have experience governing in coalition, Dutch CDA had co-governed with parties capturing over 60 percent of the non-CDA votes in 2010. Similarly, Germany’s proportional system has seen multiple “Grand Coalition” governments, as well as other multiparty coalitions, showcasing the differences between winner-take-all and proportional electoral systems. Politics in these countries is much less zero-sum than the United States as “us” and “them” shifts from election to election. As Figure 3 shows, this relationship between district magnitude and proportionality of electoral system generalizes to our 20 western democracies.

“Politics in these countries is much less zero-sum than the United States as ‘us’ and ‘them’ shifts from election to election.”

To test our expectations, we analyze data from 77 election surveys across 19 Western democracies since the mid-1990s. We utilize the widely used “feeling thermometers,” which ask respondents to rank how warmly they feel about a given party from “0” to “10.” We find that partisans evaluate former coalition partners more warmly, controlling for both the past and present ideological positions taken by these parties. The effect size for past coalition status is roughly one unit on the 0 to 10 feeling thermometer, depending on the specific model specification and we detect effects even for coalitions 10 to 15 years in the past. In one specification, we include party-pair fixed effects to show that as the coalition histories of the same two parties evolve, so do their partisan’s evaluation.

Summing up, we find that more proportional systems have denser networks of current and past co-governance, leading to warmer out-party evaluations across Western publics. Proportional representation fosters party systems with rich coalition histories, contributing to the warmer cross-party evaluations observed in these systems. This aligns with arguments for a more proportional electoral system in the United States.