Representative Don Beyer of Virginia has proposed a plan for ranked-choice voting that would make the U.S. House less partisan and more representative for all.

Imagine being a Democrat this November, the morning after this year’s midterm elections. And imagine your heart sinking: Republican gerrymanders held well in key states, and despite making more than a dozen gains, Democrats have failed to take the House.

It’s not the most likely scenario. But if Democrats remain in the minority for the fifth House term in a row, calls for wide-ranging electoral reform could become more prominent within the party. Last summer, Representative Don Beyer, a Virginia Democrat who serves an Alexandria-based district in the House, introduced the Fair Representation Act (FRA), a bill that would mandate multimember congressional districts, ban gerrymandering, and require a process in House races by which voters would indicate ranked preferences for multiple candidates in multimember districts to make the chamber more representative of people’s preferences.

Disproportionality in the House—the gap between votes won and seats held—is already striking, and the opportunity for a plurality reversal in the midterms, where one party wins the most votes and another wins the most legislative seats, is relatively high.

The Fair Representation Act emerged from talks between Beyer and FairVote, a nonprofit dedicated to electoral reform. Beyer was initially wary of the proposals, according to the FairVote president and CEO Rob Richie. “He came to this gathering pretty skeptical,” Richie says. “To his credit, [he] asked lots of hard questions.” Now, Beyer is likely the foremost evangelist for proportional representation in the United States.

When Beyer introduced the FRA last summer, joined by his co-sponsors Representative Jamie Raskin of Maryland and Representative Ro Khanna of California, he took the first step in a process that even the bill’s proponents admit is likely to be a long one. Both Beyer and Richie think the FRA is likely to become law—but they put a likely date of passage one to two decades down the line. That might come closer if feelings of disenfranchisement or frustration with unfairness in the current system continue to build, though. “If the Democrats win by a big popular-vote margin and still don’t have control of the House, it’ll spur a lot of reforms on the Democratic side,” Beyer says.

At the state and local levels, ranked-choice voting is making some inroads. Cities like Santa Fe, New Mexico, and Cambridge, Massachusetts, are on board. And, in what the FairVote legal and policy director Drew Penrose describes as a “sea change,” Maine has instituted ranked-choice voting for its elections this year, in which voters will order each candidate by preference rather than voting for only one choice. (Unlike Beyer’s plan, Maine does not use multimember districts.)

Pushing state-level efforts further will be key in building support for the measure, Penrose says. Experts like the University of Missouri political science professor Peverill Squire also note that state-level changes to the electoral system have a more immediate chance of becoming reality than federal-level shifts.

Beyer believes that proportional representation—the idea that political parties deserve legislative representation proportional to their share of the vote—would incentivize bipartisanship and reduce polarization. Rather than face primary challenges from more extreme members of their own parties, politicians would be well advised to seek out opposition votes. In a proportional, multimember district, for example, a Democrat who could win second or third preferences from Republican-leaning voters would stand a greater chance of being elected than one who drove off moderates by taking extreme stances. The system could also lead to a third-party presence in the House, raising the prospect of coalitions and a new sort of dealmaking.

So how would it work in practice?

Take North Carolina. FairVote would split the state into three districts, two with five House members, one with three. They estimate the map would elect five Democrats and five Republicans, with each district also having a swing seat that would likely vary election to election.

And that’s just if no additional parties grew or emerged. There would be no real need for Democrats to choose between a Clinton wing and a Sanders wing—they could vote for both in ranked order. A Trump critic like the outgoing Representative Mark Sanford of South Carolina or Senator Jeff Flake of Arizona could also create a viable new center-right movement. Lee Drutman, a senior fellow at New America, has argued that moderates and conservatives who are increasingly shut out of the Republican Party ought to support proportional systems since they could potentially give rise to a new party that places a greater emphasis on non-Trumpian Republican values.

In North Carolina, one five-member district could be built from roughly the area between Raleigh and Winston-Salem, essentially a combination of the Research Triangle and the Piedmont Triad. It would take in heavily Democratic college towns and major cities alongside Republican-leaning suburbs and rural areas in the state’s center and along the Virginia border. Voters would likely choose between a slate of five candidates for each major party, plus third-party candidates and independents. They’d be asked to rank each of the candidates in order of preference.

Then polls would close and the tabulation process would begin. A quota would be set for the minimum number of votes needed to be assured election. In a one-member seat, that’s just half the total plus one. For a two-member seat, it’s one-third plus one, and so on. For the hypothetical five-member district, it would be one-sixth of the total vote plus one. When candidates are elected with surplus votes, voters’ preferences are distributed to other candidates until they reach the quota for election. Through this process, a party with 40 percent of the vote in one of the new five-member districts would win two of five seats, in contrast to the current single-member, winner-take-all elections, in which all five of those seats would likely go to one party and the 40 percent of the electorate favoring the other party would have no representation.

The most notable use of ranked-choice voting in multimember districts at present is in Ireland, which uses the system for its parliamentary elections in a way almost identical to that proposed by Beyer and FairVote. According to Richie, ranked-choice voting is not the most perfect proportional system, but it is the most American of the available options because it is candidate-based rather than party-based. This would, he says, fit well with the American electoral psyche and the desire, even in an increasingly partisan era, to put candidates before parties.

Securing better representation for minority groups is one of the major arguments advanced by FairVote and Beyer. Indeed, Cynthia Terrell of RepresentWomen was a key voice in the early stages of the FRA’s creation, when Beyer was being brought on board to introduce it in Congress, Penrose says. And there are strong arguments that multimember districts lead to more representative legislatures.

Richie says that increasing minority representation—and forcing politicians to create broad appeals across partisan lines in the process—is one of the most significant benefits of ranked-choice voting. Richie co-founded FairVote in 1992, during an age of party reorientation and polarization, and now believes that within another two decades, ranked-choice voting will become the law of the land, partly because of its prospects for creating a more demographically representative House.

Across the former Confederacy, Richie says, almost every multimember district would be almost assured of electing at least one black and at least one white representative. Those members would have the same constituents and would need to appeal across demographic lines, but would also be more able to represent communities that currently don’t have much political say-so thanks to the lower threshold required for election.

The paradox of electoral reform? It’s usually only possible if you’ve already won an election. To advance proportional representation, its advocates will first need to win elections under the first-past-the-post system, and the impetus for electoral reform is often strongest in the losers of elections, not the winners who have just succeeded under the current system. And many of those elected under such a system would find their electoral prospects diminished if their single-member districts were eliminated. If Democrats retake the House, the incentive to push for electoral reform may be lessened. Beyer believes that the push for ranked-choice voting could take 10 to 20 years—and that not all Democrats who have been elected under the current system are going to be enthusiastic. “The hard part is,” he says, “the leadership of both parties has risen … in the existing system. There’s a necessary inertia in moving away from something that has worked for you.”

And then there are the issues involved with advocating for the system in a way that appeals to voters. Squire, the University of Missouri professor, says that a key problem for advocates of election reform is the tendency of voters to reject systems that are confusing or present newfangled concepts. The more complex electoral systems are, “the more difficult they are to sell to voters,” he says.

To Bruce Ackerman, a constitutional scholar at Yale University, a proportional system like ranked-choice voting could have unintended consequences. Ackerman says that by creating a multiparty system, proportional representation might strengthen an already too-strong presidency, weakening the legislative branch through increased divisions in the House. “We want a two-party system with a genuine responsiveness to regional and federal concerns and national concerns by congressmen,” he says. However, to those like Drutman for whom a multiparty system is one of the advantages of ranked-choice voting, such a critique is less significant.

The conservative writer Noah Rothman argued in Commentary last month that Democratic losses in November would be likely to provoke an increased skepticism of the constitutionally ordained process of American democracy. It is certainly possible that a plurality reversal in the fall could equate to a stronger push for reform, but measures like ranked-choice voting and multimember districts would hardly fall beyond the pall of the Constitution. Indeed, unlike other reforms, ranked-choice voting could be congressionally mandated with no need for constitutional changes.

Before the last presidential election, the then-candidate Trump said that the vote might be “rigged.” If they lose in November, Democrats could feel much the same. And they might turn to the Fair Representation Act—and other measures—as one way to fix a system that causes the winners of the popular vote to lose elections.

PARKER RICHARDS is an editorial fellow at The Atlantic.