Most elections in the United States instead use a winner-take-all system: Instead of multiple parties earning seats in proportion to their support, each election is won by a single party, usually in a single-member district. Scholarship finds that winner-take-all elections contribute to several concerning problems. Elections can be overwhelmingly uncompetitive, partisan and racial minorities are systematically underrepresented, and polarization and anti-democratic extremism tend to be more severe.
Various scholars, lawmakers, and civil society groups have therefore begun advocating that U.S. elections adopt proportional systems in lieu of winner-take-all. While proportional systems vary considerably in their designs, research consistently finds that they produce more competitive elections, more equal representation for all voters, and less pernicious polarization and associated extremism. No one model of electoral system is a panacea, but proportional systems appear to contribute to more representative and resilient democracies.
Problem Statement
How can we address the ways that winner-take-all elections harm our politics, prevent meaningful representation for certain communities, and threaten our democracy?
Background: The impact of winner-take-all elections
Legislative elections in the United States almost uniformly follow a winner-take-all rule. For the House of Representatives, each congressional district elects a single member, who “takes all” by virtue of being the only one elected. This is in contrast to proportional systems, where each district elects several members, with each party earning seats in proportion to its share of the vote. The Constitution does not require winner-take-all elections; the exclusive use of single-member districts for the House of Representatives is a policy choice mandated by federal law since 1967. That policy choice has produced a number of concerning consequences.
First, congressional elections are overwhelmingly uncompetitive. Most districts are easily identifiable as “red” or “blue,” before any votes are cast, turning the general election into a mere formality where the dominant party nominee wins by a commanding margin. As of writing, the 2024 general election is over six months away, yet Republicans have, in a sense, already won 192 seats (44 percent of the House), and Democrats have already won 173 (40 percent of the House). In these safe districts, voters will only have the opportunity to cast a powerless vote in an election that was decided years ago by cartographers. Meanwhile, control of the House will be decided by the small minority of voters (as few as 10% by some measures) who happen to live in swing districts.
In theory, congressional elections could be made more competitive by emphasizing competition when adopting district maps. However, people often live in communities that share political interests and partisan preferences, and so doing so would require an affirmative effort to emphasize competition over other considerations. Due to such geographic sorting, it will usually be easier to draw uncompetitive districts, and infeasible to draw every district as competitive. Meanwhile, most states still put the right to redistrict in the hands of partisan actors, who are often averse to emphasizing competition and instead prone to intentionally gerrymander districts. Though the effort towards independent redistricting has been valiant, and surprisingly successful, as long as single-member districting is used, gerrymandering will exist. Proportional, multi-member districts are extremely hard to gerrymander, while winner-take-all, single-member districts are uniquely prone to gerrymandering.
Second, single-member districts exaggerate the electoral power of majorities and underrepresent minorities. Because of their “take-all” structure, a party that wins more than 50 percent in a district (or in many cases, only a plurality) effectively wins 100 percent of representation. When patterned across a state, this dynamic can produce significantly biased results. For instance, in Massachusetts, despite commanding roughly a third of the statewide vote, Republicans regularly win zero of the state’s nine House seats. In Oklahoma, Democrats likewise win roughly a third of the statewide vote, and yet win zero of the state’s five House seats.
These electoral biases carry especially troubling implications for racial minorities, who tend to likewise be underrepresented in winner-take-all elections. Where racially polarized voting exists, the only way a minority group can elect a candidate of choice is by being the largest group of voters in a district. The result is that racial minorities, even if a substantial voting bloc, can be systematically underrepresented. In Alabama, Louisiana, and South Carolina, for example, Black voters constitute over 25 percent of voters overall, yet candidates of choice of the Black community win roughly 15 percent of their House seats.
To partially correct for these outcomes, states may deliberately design “majority-minority” or “minority opportunity” districts. Absent a minority opportunity district, Black voters would likely elect no representatives of their choice in South Carolina, for example. At its best, this approach is limited: it relies on patterns of residential segregation that may not (and arguably should not) exist for groups in the near-future, and does not exist for some racial groups now. Even where districts can be effective, the approach of achieving minority representation through redistricting relies on a legal requirement for such districts to be drawn, given that the map-making authority is often majority-controlled. That legal requirement is found today in the Voting Rights Act, but its dictates have been significantly weakened by judicial decisionmaking.
Third, winner-take-all elections are associated with higher levels of affective polarization. When elections are winner-take-all, politics drives toward a binary: Either “we” win and “they” lose, or “they” win and “we” lose. This creates a zero-sum dynamic where partisans can gain just as much by denigrating and alienating their opposition as by making a positive case for themselves. Internationally, this kind of negative, “affective” polarization is most severe in countries with winner-take-all elections. In countries with deep and severe division, such as the United States, winner-take-all elections tend to exacerbate divisions, whereas proportional systems create mechanisms that facilitate greater management of partisan conflict.
All of the above dynamics combine to make our system vulnerable to authoritarianism. Distorted partisan representation can result in one party winning majority control without majority support among the electorate. Such a distortion creates strong incentives for it to further frustrate majority rule and entrench its advantage. That includes eroding the legal protections for enfranchisement of minority voters: both in terms of blocking access to voting (voter suppression) and offering only a meaningless ballot to people of color (vote dilution). With few competitive districts, the ability of voters to push back against an increasingly authoritarian party (by voting for its opposition) is severely diminished. As partisan polarization becomes more and more severe, authoritarians can gain support from otherwise pro-democracy voters, because the only available alternative is voting for a party that feels not only opposed on policy but alien in identity. This all combines to create a “doom loop” that threatens democracy itself.
Proposed Solution: Proportional representation
To address the problems attributable to winner-take-all elections, the U.S. should join the majority of other modern democracies in adopting a system of proportional representation.
Proportional representation is a class of electoral system that aims to ensure seats reflect votes: that is, a political party’s share of seats in a delegation and a legislature is commensurate with its share of its vote: If a party wins 30 percent of votes, it should earn 30 percent of seats, and likewise for any other share. Proportionality is achieved by using multi-member districts instead of single-member districts, in which multiple seats are allocated according to parties’ vote shares. For example, a state like Louisiana that elects six members to the House of Representatives could elect all six statewide on a proportional basis, or it could divide into two equal-population districts that each elect three members on a proportional basis.
Proportional representation does not guarantee any particular outcome for any groups. Rather, it makes it possible for a share of voters to elect an equal share of seats. The parties and candidates must still make the case to voters and earn their votes. But any party that earns those votes will also win its fair share of seats.
While all proportional systems aim to ensure proportionality in outcomes, no two systems around the world are exactly alike. When countries adopt proportional systems, they do so within the context of their political histories and with a mind to their specific challenges. While there are certain common models of proportional representation, details vary considerably. For example, multi-member districts could elect as few as three winners or as many as 10 or more ( “district magnitude”). Within a multi-member district, voters could vote for a single candidate as they do today (as in “open list” proportional representation), or they could vote only for a political party (“closed list”), or they could rank candidates (“proportional ranked choice voting”). This Working Paper does not advocate for any specific system, and indeed given the federal nature of the United States, different states could adopt different systems, allowing for local variation and experimentation.
Electoral competition
With proportional representation, there are no “red” or “blue” districts. The greater the district magnitude, the easier it is for a party to win a seat, and the smaller a percent of the vote is needed to change the outcome. In other words, proportional multi-member districts will be nearly universally competitive: No district would be totally safe for one party.
Proportional systems are less vulnerable to both geographic sorting and intentional gerrymandering. Republicans in cities and Democrats in rural areas would be expected to win a share of seats commensurate with their support, rather than being completely shut out. In Massachusetts, Democrats would still win a majority of seats, but Republicans would win their fair share as well. Likewise, in Oklahoma, Republicans would still win most of the seats, but Democrats would earn representation too.
A higher district magnitude makes attempts to gerrymander districts for partisan advantage increasingly difficult and less effective. Internationally, we see far less susceptibility to gerrymandering in countries with proportional systems. Even a modest increase in district magnitude can significantly blunt attempts to gerrymander, and gerrymandering becomes functionally impossible where districts each elect at least five members.
Minority voting power
Minority groups, including women and communities of color, have battled disenfranchisement and attempts at voter suppression throughout much of American history. Proportional representation is one way to further the U.S.’s progress in representing historically marginalized communities.
Studies show that winner-take-all systems tend to systematically underrepresent racial minorities, and that proportional systems are more likely to provide for both greater descriptive and substantive minority representation. Under winner-take-all, the dominant voter-group “takes all” in a district; and so unless a racial group can garner support from a majority (either by being the majority themselves or by convincing enough non-minority voters to support their preferred candidates), they remain barred from representation. Such is the case for many Black, Latino, Asian, and Native voters, particularly those who are geographically dispersed outside of minority opportunity districts. In Nevada, for example, Latino voters comprise nearly 20 percent of the electorate, but there are no Latino representatives in the House delegation.
A majority of Black voters in southern states with a history of racially polarized voting live in majority-white (often safe Republican) districts. Modeling exercises find that if these states were to utilize proportional multi-member districts in the place of the single-member districts, the share of Black voters able to elect at least one candidate of their choice increases to 98 percent. Comparative research consistently finds better representation for ethnic minorities in countries that use proportional representation.
More people of color would be able to elect a candidate of their choice because proportional representation would translate votes to seats proportionally, liberating the power of any given group of votes from the need for a favorable district map to secure representation. For example, in a proportional multi-member district with three seats in which Black voters constitute one third of the population, they can be expected to elect their preferred candidate to at least one seat if they vote cohesively (and more seats if they garner support from non-Black voters for additional candidates of choice).
If this same region is divided into three single-member districts, however, the ability of Black voters to elect their preferred candidate becomes dependent on their geographic distribution and on the choice of district map. To meet the needs of its diverse population, the United States should consider an electoral design which utilizes district magnitudes of five or more to minimize the potential for racial vote dilution and to enable representation of those members of the electorate who do not constitute a majority in their district. Such a design would make the Voting Rights Act’s guarantee of an equal opportunity to elect harder to violate through vote dilution.
Proportional representation may have particular significance for women’s representation as well. In 2023, women were 50.4 percent of the overall population, yet women representatives comprised only 28 percent of the U.S. House. While there is rarely parity between rates of female and male representation in democracies, the U.S. performs particularly poorly when compared to countries that elect proportionally, such as Germany (35%), Denmark (40%), or New Zealand (50%). Overall, countries that elect proportionally elect more women. Female candidates also tend to be more successful when running in multiparty elections and in multi-member districts.
Fairness of partisan results
Proportional representation can also almost guarantee partisan fairness in electoral results. With single-member districts, the choice of district boundaries can profoundly change who wins and loses with any given group of voters, whereas with proportional representation, a share of votes elects a roughly equal share of seats. Taking Oklahoma as an example: In 2022, Democrats won between 22 and 37 percent of the vote in each of its five congressional districts (for 31 percent of the vote statewide), winning zero seats. This large proportion of wasted votes in single-member districts can create severe distortions in aggregate. For example, in 2012, Republicans were able to win a majority of seats in the House of Representatives despite Democrats earning significantly more votes in aggregate.
Proportional representation would essentially eliminate these disparities, both within states and in aggregate. Returning to Oklahoma, a proportional formula would ensure that the Democratic minority within the state would have the votes to elect one or two of the five representatives. In fact, simulations suggest that even with modest district magnitudes, no multi-member district in any state would be likely to exclusively elect a single party slate; Democrats would win in places like Oklahoma and Republicans would win in places like Massachusetts. When aggregated across all districts, proportional representation results in a much closer fit between votes and seats.
Correcting for the severe partisan imbalances of single-member districts is often the most straightforward case for proportional representation. For example, when New Zealand elected its national legislature in single-member districts, it suffered two consecutive elections in which one party won a majority of seats despite its opposition earning more votes. This immediately led to a push for proportional representation that ultimately succeeded in 1993, leading to the country’s use of a proportional electoral system ever since (with no such electoral inversions occurring since its adoption).
Polarization and extremism
Proportional representation, by ending the all-or-nothing, binary choice associated with winner-take-all, would help reduce partisan polarization. In general, countries with more proportional systems tend to have less affective polarization, tend to be more effective at managing political conflict, and tend to have lower levels of ethnic-based political violence.
In a single-member district, demonizing the opposition is often an effective way to win elections. In districts that are safe for one party, the election is effectively decided at the primary election stage, where a smaller, more partisan, and less representative group of voters decides the winner. In competitive districts, hostile negative campaigning can drive down turnout for an opponent. These incentives are reversed under proportional representation, where there are no safe districts, and where the greater number of choices means that campaigning is less zero-sum - attacking an opponent does not necessarily help when the voter has other choices available.
More broadly, proportional representation can facilitate multiparty politics, which tends to result in shifting coalitions, “where few political enemies are ever permanent.” In countries with proportional representation, voters tend to view opposition parties less negatively. Likewise, opposition parties themselves are more likely to accept the results of elections under proportional representation, in part because rather than being shut out, they will have earned their fair share of political power and influence. In this environment, politics may still be quite partisan, but will not develop patterns of affective polarization that can lead to extremism and democratic decline.
The way that proportional representation addresses these issues is particularly important. Historically, conventional wisdom held that the two-party system prevented extremism by suppressing extremist parties. However, that very dynamic created a vulnerability: if extremists could capture one of the two major parties, they could significantly expand their influence. Under proportional representation, on the other hand, extremists can win some seats, but never more than the share of votes they are able to earn. For example, Poland’s 2023 elections show how an otherwise-unlikely coalition of pro-democracy parties can overcome a rise in authoritarianism under a more proportional system.
Next Steps: How we can move toward proportional representation
Winner-take-all elections are not inevitable. They are a policy choice. Already, reformers and voting rights activists have made progress in moving towards more proportional systems.
There is a history of advocacy for proportional representation at the state and local levels. From 1870 to 1980, Illinois elected its state house of representatives in three-member districts using cumulative voting, a semi-proportional system. Because cumulative voting is only semi-proportional, and because districts only elected three winners, the system was modest in its impact - the state house still had many uncompetitive contests and a two-party system. However, even such a modest system had a significant impact. Both Republicans and Democrats won seats in all parts of the state, resulting in a less polarized and more effective state legislature during its use, and Black members won election to the House throughout the Jim Crow era when we saw very few Black state legislators elected in southern legislatures.
From 1915 through the 1940s, some 24 cities, including Cincinnati and New York City, experimented with proportional ranked choice voting (Cambridge, Massachusetts retains the system from this era). These uses have a record of improving representation. For example Cincinnati elected its first Black councilmembers under the system at a time when the city was about 10 percent Black. The one city that used the system in partisan elections, New York City, went from a two party system to a multiparty system (then back to a two party system after returning to winner-take-all).
In the 1990s, some of the first tentative steps toward greater use of proportional representation came as the result of litigation brought under the Voting Rights Act. Vote dilution lawsuits brought under Section 2 of the Voting Rights Act usually seek to replace winner-take-all at-large elections with districts, including some number of majority-minority or opportunity districts. However, some of these cases have instead ended with the parties agreeing to a “modified at-large,” which retains at-large elections but replaces the winner-take-all voting rule with something more proportional.
To date, more than 100 local jurisdictions have adopted a semi-proportional method, limited voting or cumulative voting, to remedy a vote dilution claim. Although these methods are far from proportional, they do make it possible for a minority bloc of voters to elect a candidate without the use of districts. A few recent cases have even adopted proportional ranked choice voting, a more truly proportional system. The use of proportional representation can sometimes be a more effective remedy than districts, and it can also allow the jurisdiction to retain the at-large nature of the election, which it may prefer in some contexts.
In the wake of the Voting Rights Act being successively weakened by federal courts (see above), some states have passed state-level Voting Rights Acts, and these often explicitly name semi-proportional and proportional voting methods as available remedies in vote dilution lawsuits.
Outside of the litigation contexts, advocates have begun a new wave of momentum for proportional representation at the state and local levels. The diversity of places considering the reform speaks to its many benefits: For example (deep blue) Portland, Oregon became the first new city to adopt proportional representation for its city council in 2022 and will use it for the first time in November, 2024. While, the Rainey Center’s Andy Craig recently published a report arguing for proportional representation for the (ruby red) state legislature of Wyoming.
For Congress, proportional representation, while certainly a big change, may not be quite as big a lift as some may assume. The exclusive use of single-member districts to elect the House of Representatives is nowhere to be found in the Constitution. Rather, it comes from the Uniform Congressional District Act, a law passed in 1967. States could start using proportional representation to elect their congressional delegations if that law were simply amended to make that possible. Amending the UCDA would cost nothing and would not require any changes to take place that states did not want to adopt voluntarily.
Most elections in the United States follow a winner-take-all rule, but this is a choice that we can and should revisit. Adopting proportional representation would carry benefits that speak directly to many of our most pressing political problems. Consideration of proportional representation should be a high priority for those interested in improving our politics.