Polarization is severe

The United States is in a dangerous place politically. Two-party political polarization has made effective governance all but impossible and threatens serious democratic decline. The evidence is clear: Since the 1970s, Republicans and Democrats in Congress have voted together less and less. Over the same period, voters have become more reliably partisan, with fewer voters “splitting the ticket” (voting for Republicans for some offices and Democrats for others). 

Voters have also become more geographically and culturally sorted: Urban voters are becoming more Democratic and rural voters more Republican. Notably, voters are not becoming much more extreme or polarized around policies; many Republicans believe in raising income taxes for wealthy households, and many Democrats believe government spending is often wasteful.[1] Nonetheless, voters increasingly see one of the two major parties as a threat — often an existential one.

In April 2013, leading election law scholar Rick Hasen published a paper arguing that congressional polarization has made governance unacceptably dysfunctional. Unless it self-corrected, he warned, constitutional change would be necessary.[2] Far from self-correcting, polarization has worsened in the years since. In 2015, political journalist Matt Yglesias wrote a piece with a franker title: “American Democracy is Doomed.” In 2017, political scientist Lee Drutman echoed the point in an article arguing that polarization could not self-correct because we had entered a self-reinforcing cycle of escalation, which he ominously called a “doom loop.” In 2020, Drutman expanded on this thesis in his book Breaking the Two-Party Doom Loop: The Case for Multiparty Democracy in America. In it, he argues that to break the cycle we need to replace winner-take-all elections with a proportional system.

These warnings showed great foresight. The Trump years featured several instances of political violence, culminating on January 6, 2021, with a mob storming the U.S. Capitol building  while Congress counted the electoral votes inside. The insurrection resulted in five deaths, including that of a police officer, and several nonfatal casualties. This riot was incited not only by Trump’s false and unjustified insistence that he won the 2020 election, but also by the many Republican members of Congress and other political elites who legitimized what has come to be known as the “big lie.” 

Much of the contemporaneous analysis of January 6 blames only its most proximate instigators: Trump, conservative media, contributing lawmakers, and the Republican Party. However, much of the elite support for Trump’s baseless claims came in response to electoral incentives, as explained more fully below. The problem is not merely individual; it is institutional. It cannot be resolved merely by holding bad actors accountable. We must also reform our institutions.

Partisanship dominates voting behavior

Not too long ago, “ticket splitting” was a regular phenomenon in many parts of the country. Many Americans voted for the Republican nominee for president and the Democratic nominee for their congressional district, or vice versa. Once in office, members formed coalitions across party lines as well as within them. Arguably, we were closer to a four-party system than a two-party system, because each party had its own liberal and conservative wings. Members negotiated for earmarked federal spending in their districts in exchange for their votes, and often campaigned not on divisive national issues but on their records of acquiring federal dollars to benefit their constituents.

Though productive, this period reinforced racial inequality and oppression. It does not offer a model for the future. Bipartisan policymaking relied on the tacit acceptance of racist Jim Crow policies throughout the South. Prior to the “Republican Revolution” in 1994, the Democratic Party held what appeared to be a permanent majority in the U.S. House because its coalition included Southern Democrats who opposed civil rights and racial equality. 

The foundations that made this era of bipartisanship possible were unstable and unacceptable. It is now long over. Many changing features of American politics led to its demise, including partisan sorting over civil rights, the decline of local news media, a ban on earmarks in federal legislation, and more. But the incentives inherent in winner-take-all elections made its decline a one-way ratchet, toward greater and greater hostility and with instability and violence as the only endpoint in sight.

Our new true two-party system can be clearly seen in the alignment of presidential election results with congressional election results. The presidential election is quintessentially national. If voters generally favor the same party for both Congress and the president, then congressional elections are nationalizing. That is, voters care less about who represents them than about which party holds majority control of the U.S. House.

Historically, the chamber’s most moderate members came from districts that favored the other major party. These members kept their jobs because they were able to connect with voters and colleagues across the aisle. Democratic Rep. Colin Peterson, for example, represented a heavily Republican district in Minnesota from 1990 to 2020. He won election each cycle, even as his district was carried by Republican presidential nominees George W. Bush, John McCain, Mitt Romney, and Donald Trump. 

In 2016, Trump won Peterson’s district by a stunning 30-point margin, even as Peterson beat his Republican challenger 53-47. Peterson earned support from Republican voters in part by maintaining a moderate voting record in Congress throughout his three-decade tenure. But that is not enough today. After 15 terms, he finally lost to a Republican in 2020. 

That year, the winning congressional candidate came from the same party as the presidential candidate who won the district in 419 out of 435 districts. No candidate won in a district that favored the other major party’s presidential nominee by more than an 11-point margin.[3]

National partisanship and the nationally favored party have been dominant factors in congressional elections for decades, but, since the 2020 elections, they have essentially become the only factors. As a result, today’s candidates win congressional general elections by winning the primary election for the party favored in the district that year. Everything else matters only at the margins.

Winner-take-all drives polarization

Nationalization of politics is not inherently bad, but, when applied to a two-party system with 435 separate winner-take-all elections, it creates strong incentives for the two parties to sharply polarize. 

The use of winner-take-all elections in a nationalized environment shapes the incentives faced by incumbents who want to be reelected. Rather than campaigning on a record of serving the district, representatives win over voters by campaigning on how terrible the other major party is. That strategy is particularly important for winning primary elections, with their smaller and more ideological voter bases. As a result, negative partisanship has come to dominate American politics. Very few people like either of the two major parties all that much, and people increasingly hate and/or fear one of them.[4]

On the flip side, incumbents do the most harm to their reelection chances when they cooperate with the other major party. Consequently, legislative records show that representatives are increasingly inclined to vote with their parties. Under a measure known as DW-NOMINATE, representatives earn a score of -1 if they vote with Democrats 100 percent of the time and a score of +1 if they vote with Republicans 100 percent of the time. All members fall somewhere between these two values, but they are clustering more closely together in two disparate poles. The 112th Congress, elected in 2010, was the first to feature no partisan overlap on the DW-NOMINATE scale. During that two-year cycle, the Democrat most likely to vote with Republicans  voted with Democrats more than any Republicans did (and vice versa).[5] This phenomenon has held true in every cycle since.

DW-NOMINATE clustering over time

Source: FairVote, Fair Representation Act Report

Like nationalized politics, party-line voting is not necessarily bad. In countries with parliamentary systems, voters usually vote based on national partisanship, and elected members vote with their party virtually all of the time. After each election, the parties form a majority government and enact their party programs.

But the United States does not have a parliamentary system. Instead, we have a complex system of checks and balances that prioritizes compromise. When compromise becomes impossible, it leads to severe legislative gridlock. We see this in periodic government shutdowns when control of the U.S. House, U.S. Senate, and presidency is divided between Republicans and Democrats (as when Republicans control the Senate but Democrats control the House).

Unchecked polarization causes many more problems than gridlock, including democratic decline. As political scientists Stephan Haggard and Robert Kaufman describe in their book Backsliding: Democratic Regress in the Contemporary World, a sharply divided electorate tends to prioritize victory for its side over procedural values like a commitment to democracy, while patterns of gridlock and policy swings reduce trust in government and other institutions. When these trends co-occur, it creates an opening for autocrats to win elections and use their victories to undermine the democratic process. Haggard and Kaufman illustrate this pattern with real-world examples of democratic backsliding in Brazil, Poland, and the United States.

Polarization in the United States has already eroded our democracy. The public still expresses strong support for liberal democracy over authoritarianism and generally prefers pro-democracy candidates, but voters are increasingly willing to sacrifice fair process to help their side win. 

Recent research by political scientists Matthew Graham and Milan Svolik (both at Yale University) reveals troubling trends in how voters react to candidates who violate democratic principles.[6] If voters are faced with a candidate from their preferred party who violates fundamental democratic principles, they are unlikely to break ranks and vote for an opposing candidate. Because only a small percent of centrist voters do that, candidates who openly flout democratic values face minimal punishment. Graham and Svolik also find that increasing the ideological distance between party platforms makes voters even less likely to break ranks.

The more voters view the other major party as a threat, the more they will accept anti-democratic positions over a victory by the other party. That creates incentives for political hardball, where representatives ignore established norms that promote compromise and functional governance. Prior norms of restraint in the use of the filibuster, executive orders, opposition to presidential nominations, and so on, have broken down in recent years. When one party breaks a norm (or is perceived to do so), the other major party has little choice but to reciprocate, lest it be accused of unilateral disarmament. This creates a reinforcing cycle, and there is only so much escalation that can occur before violence breaks out.

Merely “restoring” depolarized politics cannot be the goal

Only one period in U.S. history was more politically polarized than the present: the years leading to and during the Civil War. In 1860, Abraham Lincoln was elected the nation’s 16th president, despite losing every Southern state. The new Republican Party that nominated him was founded by a coalition of former Whigs and Democrats who opposed expanding slavery in new territories. 

Lincoln campaigned as a moderate, and explicitly promised not to interfere with slavery in the states where it was already practiced, but the Republican victory was nonetheless seen as an existential threat to Southern whites. They made the radical decision to secede from the Union, leading to the Civil War.

The North, of course, won, abolished slavery, and restored the Union, but the nation remained dangerously polarized during Reconstruction, when the federal government sought to reintegrate the formerly Confederate states into the Union. Polarization eventually receded, but not by reconciliation, the passage of time, or integration of freed Black Southerners into political power. It receded rather by Northern Republicans’ acquiescence to Southern Democratic demands to end Reconstruction and permit the total suppression of Black political power in the South.[7] Once the issue of racial equality was indefinitely tabled, compromise became possible again, allowing Congress to operate more productively.[8]

The United States has never truly achieved a stable multi-racial democracy. When the issue of racial equality predominates, sharp polarization results; throughout history, our only response has been war or overlooking racial oppression. Restoring a more productive politics by tolerating racial oppression would be unconscionable, even if it were possible. Instead, we must find a way to mediate our differences – to peacefully share political power among people of all races and backgrounds.

Unfortunately, winner-take-all voting rules in a two-party system make peaceful resolution of our conflicts prohibitively difficult because they distill our differences down to two “sides.” Effective, compromise-oriented governance relies on overlapping, cross-cutting coalitions. If the two parties represent entirely distinct worldviews, such that each views the other as threatening to core identities, then opportunities for compromise will be replaced by the accelerating “doom loop” of negative partisanship that we see today. To break that loop, we need a system that allows for the election of candidates who prioritize inclusive democracy over ideological fervor and genuine problem-solving over obstruction. That is only possible outside the confines of a winner-take-all, all or nothing, binary choice.

Mediating our differences

The kinds of cross-cutting values and interests that make compromise possible do exist. In 2021, the Pew Research Center updated its 2017 research on political viewpoints.[9] The new report confirmed that partisans view their partisan opponents with ever-increasing hostility, but it also showed that, when it comes to policy, voters are not strictly red and blue. On even the most divisive issues, about a quarter of partisans hold viewpoints associated with the other major party – for example, about a quarter of Republican-leaning voters say we should raise the minimum wage, and about a quarter of Democrats say that government is almost always wasteful and inefficient.[10] This graph, from a November 2021 paper by political scientist Alan Abramowitz, shows the degree of modern overlap on left-right issues.[11]

Distribution of Democratic and Republican identifiers on Left-Right issues

The overlap is shrinking, but it is far broader than the simple polarization narrative would suggest. Many fiscally conservative voters still believe in expanding opportunities for immigration, many socially liberal voters want fewer regulations on business, and so on. However, zero-sum, winner-take-all congressional elections do not provide these voters with the opportunity to elect like-minded representatives.

In our winner-take-all elections, we instead face a choice between one Democrat and one Republican (if the election is competitive at all). These nominees may reflect the views of the median Democrat and median Republican, but they wash out the great diversity of political opinions among all Democrats and all Republicans, not to mention unaffiliated and third-party voters. The median partisan Democratic or Republican voter not only lacks a mix of liberal and conservative viewpoints, but also tends to view the other major party extremely negatively. Such extreme partisanship does not reflect the views of most voters, let alone most people. But in a head-to-head contest where only one candidate can win, it dominates our political choices.

When only one group of voters can elect one candidate, elections gravitate toward the sort of competing worldviews that create and accelerate polarization. To break this cycle, we need a voting method in which multiple candidates win, each representing distinct viewpoints. Such a method will create space for candidates to earn their seats by campaigning to groups outside of those that hold the median partisan perspectives. Once in office, these candidates can then form coalitions either within or across party lines in the interest of their constituents. This is precisely what proportional representation can accomplish.