Two weeks before Maine’s June primary—the first statewide election held using ranked choice voting (RCV)—David Brooks penned an opinion column in The New York Times headlined, “One Reform to Save America.” Brooks argued that ranked choice voting, particularly in multi-member districts, could cure the evils that have led to a polarized two-party system in America. Beware a reform that promises too much.
At least since French political scientist Maurice Duverger published Political Parties more than seventy-five years ago,1 political scientists have preached that single member districts lead to two-party systems. My own research with John Bibby, exploring why the United States maintains a two-party system despite widespread dissatisfaction with the choices it presents, similarly emphasizes institutional factors as the driving force behind two-party competition.2
But would altering election rules really encourage minor party candidacies or end the polarization in our current two-party system? Is ranked choice voting—in which candidates are incentivized to compete for second and third-choice support from a politically diverse electorate—the best way to achieve these ends? And would the end result be worth it?
Who came up with ranked choice voting—and why
The English pamphleteer Thomas Hare first proposed ranked choice voting in the mid-19th century as a way to accommodate minority interests while guaranteeing majority rule.3 The single transferable vote (STV) system used in Ireland and other countries is essentially an RCV system used for districts with multiple seats. To be elected, candidates must reach a minimum number of votes known as a quota. Voters cast an initial ballot for their first-choice candidate, then, once one candidate has achieved the specified minimum, any surplus votes that candidate received are transferred to each voter’s second-favorite candidate, then third-favorite, and so on. The process continues like that until all the seats have been filled; if no single candidate reaches their quota, the least popular candidate is eliminated and their votes are again redistributed to each voter’s second-choice candidate. Many argue that STV is preferable to straight proportional representation in parliamentary systems, because citizens can choose individual candidates as well as their party preference.
A number of U.S cities instituted similar schemes during the Progressive Era. RCV was implemented for municipal elections in communities as small as Ashtabula, Ohio and Norris, Tennessee, as well as large metropolises like New York City and Cincinnati. In each case, reformers pushed for the change to encourage broader representation, particularly of immigrants, and in each case the establishment fought back. During the Red Scare in the aftermath of World War I and again in the McCarthy era, establishment politicians raised fears about the sorts of extreme parties that might gain seats under RCV. By the end of the Eisenhower Administration, of the 24 cites that had once chosen city council members by ranked choice voting, only Cambridge, Massachusetts continued choosing city officials via an ordinal ranking scheme.
Starting in 1992, a group of scholars, reform activists, and public officials came together, first as the Center for Voting and Democracy and now under the name FairVote, to push for electoral reforms of various types, including Electoral College reform, implementation of proportional representation, and ranked choice voting in multi-member or single-member districts. The stated mission of FairVote is to promote fair elections, defined as elections in which every vote counts and all voters are represented. They seek alternatives to winner-take-all elections and instant run-off voting as an alternative to plurality and traditional run-off elections. Since 2000, RCV has been adopted in about a dozen, mostly liberal cities, including San Francisco, Minneapolis, and Portland.
Ranked Choice Voting in Maine
Despite this grassroots popularity, however, there had not been a single statewide experiment with RCV until Maine’s gubernatorial elections this year. Maine has a long history of electing governors with a plurality of the vote but not a majority. In nine of the 11 gubernatorial elections since Maine elected independent James Longley in 1974, the winner has lacked a majority; in five of those elections, the winner polled less than 40 percent of the vote.
Ranked choice voting advocates have been introducing RCV bills to the Maine legislature ever since 2001, none of which has made it out of committee. However, after Governor Paul LePage won election to the Blaine House by fewer than 10,000 votes over Democrat-turned-independent Eliot Cutler and Democrat Libby Mitchell in 2010, earning fewer than 38 percent of the votes cast, RCV advocates decided to bypass the legislature and change the process through referendum. Thus began a soap opera in which reformers tried to circumvent the establishment and the establishment—especially the Republicans—fought back.
RCV activists began circulating petitions to get the reform on the ballot prior to the 2014 election, which LePage won, again without a majority. In November 2015, Maine’s Secretary of State Matthew Dunlap certified that the petitioners had enough signatures to place their referendum on the November 2016 ballot. Voters approved the reform by over 30,000 votes, 52 percent to 48 percent, in that election.
But that was hardly the end of the story. In February 2017 the Maine State Senate (controlled by the Republicans) asked the Maine Supreme Judicial Court for an opinion on the constitutionality of RCV as it applied to general elections for State Representative, State Senate, and Governor. In May the Justices unanimously ruled in an advisory opinion that using RCV in those elections violated the Maine constitution, which specifically calls for election by plurality. That provision had been inserted in the Maine constitution, first for the House, then the Senate, and eventually for Governor, after a series of elections in the late 19th century in which a majority was not achieved; one, in 1878, nearly led to bloodshed, and allowing plurality winners was a compromise to avoid catastrophe.4
Chastened by the court’s ruling, Maine’s state legislature passed a bill that put off implementation of ranked choice voting until 2021. But the RCV activists did not accept that solution. The Maine constitution contains a people’s veto provision that permits citizens to petition to veto a law passed by the legislature. If the petition is certified, the law does not take effect until after citizens have the opportunity to vote on the veto. In March of this year, Secretary of State Matthew Dunlap certified that the petitioners had been successful in collecting enough signatures; and, therefore, the legislation postponing the implementation of ranked choice voting until 2021 would be on the June ballot. Until that vote, the original referendum was law—and thus the 2018 primaries (which had not been considered by the Supreme Judicial Court) would be run under RCV.
The Republican majority leadership in the State Senate challenged this outcome in both state and Federal courts, and in each case, the jurists ruled that the vote should go forward. Thus the June 12th primary was the first statewide election held under ranked choice voting. This meant Maine had to come up with a plan for implementing RCV in very short order, a thankless task made all the more difficult by the legislature’s refusal to appropriate funds to help facilitate the transition. So the first statewide experiment with RCV took place under less than ideal conditions, to put it mildly.
What does RCV purport to do? What problems exist?
Proponents of ranked choice voting say the system advances basic democratic values. Seven arguments are especially popular:
- Winners under an RCV system have majority support. While the winner might not have achieved a majority on the first round of voting, after successive candidates are dropped from the counting, the eventual winner does have a majority on the last round.
- Voters are given more choices, and they can support these choices without the fear of wasted votes. Under a plurality system, voters fear that votes for their “real” first choice would be wasted and might even help the candidate they favor least. Think of the votes for Ralph Nader in Florida in 2000, votes that otherwise might have gone to Al Gore. Under RCV, if no one achieves a majority in the first round, a vote for the candidate with the fewest votes is redistributed to the candidate the voter would have favored had his or her first choice not been on the ballot. Thus, the voter cast his or her vote for his or her preferred candidate, but the vote was not wasted even though it did not help what turned out to be the least preferred candidate.
- RCV minimizes strategic voting. Because there is less potential for wasted votes, voters do not have to vote for the “lesser of two evils” as their first choice. For this reason, RCV arguably enhances freedom of conscience during contentious elections (a freedom quite a few voters might have appreciated in 2016).
- Negative campaigning is discouraged. The logic here is that candidates will not want to attack their rivals, because people are unlikely to cast their second-choice vote for the candidate who has has attacked their first choice. In other words, civility becomes a rational political tactic, whereas name-calling and attack ads are penalized.
- RCV should lead to more moderate candidates. Under a plurality winner system, candidates on the ideological fringe have the advantage because they can win by attracting only a relative small subset of the total electorate. Under RCV, by contrast, second and third choice votes can swing the election, so candidates are rewarded for courting voters outside their immediate base. This would be especially welcome in primary elections, where voter turnout tends to be more partisan and ideological than in general elections.
- RCV should lessen the impact of money in politics. The argument here is that too much money is now spent on negative advertising. Under RCV, the emphasis should be on positive campaigning, especially grassroots campaigning, reducing overall costs and limiting the role of lobbyists in elections.
- RCV saves money over traditional run–off primaries. If citizens vote only once, elections need be run only once. Thus, if a state or community has a system of run-off primaries in place, as do many southern locales, RCV would save money.
All these claims are hotly contested by opponents of ranked choice voting, who adduce a number of counterarguments:
- Under RCV, the winner does not necessarily poll a majority of the votes cast. Because of voter fatigue, voters do not always rank all of the candidates in a race. Therefore the “majority” winner on the last round of counting might not have had a true majority of all of the votes cast; the denominator used to reach 50 percent +1 has been reduced in order to reach a majority winner. (There is also a mathematically esoteric possibility that the true choice of a majority of the voters does not emerge as the winner under RCV.)
- Negative campaigning will not be reduced. Much of the negative campaigning in American politics is done by independently funded organizations, not by the candidates. These organizations do not have the same incentives as the candidates to avoid alienating their competitors’ supporters. This also refutes the claim that campaigns will cost less, since most of the money spent on ads is raised by independent organizations.
- Fringe candidates will be encouraged to run, further splintering the electorate and polarizing our polity. This argument is often made regarding multi-member districts, such as those that led to Communist Party winners in New York City in the post-World War I era, though single-member districts aren’t immune either. Even if fringe candidates don’t win, they will likely receive more support under RCV than under a plurality system, precisely because RCV mitigates the traditional concerns about wasted votes. Fringe candidates may therefore achieve greater cultural visibility, reinforcing polarization and crowding out moderate voices.
- The system is expensive to administer. While it is undoubtedly true that RCV elections are less expensive than traditional run-off elections, it is also true that they are more expensive than plurality winner elections. The debate is over how much more expensive they are.
- RCV has perhaps worked in small locales, but it has never been tried in larger constituencies. Even Ireland and Australian states are small in terms of population by American standards, as are most of the cities in which RCV is currently in use. Opponents in Maine argued that statewide implementation expanded the size of the electorate too much; others would argue that Maine is too small a state to use as a model.
- Voters will be confused and therefore not turn out to vote. Turnout, particularly in primaries and midterm elections, is already disappointingly low. If voters are confused by the system of voting, they will be even less likely to vote.
The Evidence from Maine
The experience in Maine resolved some of these debates but left others open. First, it should be recognized that the Maine experiment was with ranked choice voting in single-winner primary elections. It tells us very little about how RCV would function in primaries with multiple winners, let alone in a U.S. general election. Second, none of the candidates in the June election were particularly extreme, so the RCV-favors-moderates hypothesis remains untested.
Those caveats aside, we can still draw some important conclusions from the Maine elections. One key takeaway is that second- and third-choice votes may not matter quite as much as some reformers seem to think. In the GOP primary, Shawn Moody, a businessman supported by Governor LePage’s followers, won nearly 57 percent of the first-choice votes on June 12. Thus, while Republicans noted their second, third, and fourth choices, these votes were never tabulated; Moody won on the first round. In the Second Congressional District Democratic primary, Jared Golden won over 49 percent of the first-choice votes; his major opponent, Lucas St. Clair, won 41 percent, and Craig Olson won under 10 percent. Golden was not declared the winner on election night, but St. Clair all but conceded, noting that he would have to have been the second choice of virtually all of Olson’s voters in order to overcome Golden’s lead. When the second round of votes was tallied, Golden won with more than 54 percent of the vote.
The most interesting race was for the Democratic gubernatorial nomination. Four candidates were considered to be the pack leaders: Attorney General Janet Mills was considered the favorite throughout the campaign, but she was challenged by former Speaker of the Maine House Mark Eves, businessman and Iraq War veteran Adam Cote, and by the liberal lobbyist Betsy Sweet, who ran as a Clean Elections candidate and thus qualified for public funding. In addition, there were three other names on the ballot, though none of them received much attention. Mills led Cote by about 6000 first choice votes but was well shy of a 50 percent majority. Four rounds of counting were needed to determine the winner, but despite Cote’s hopes that he would gain votes from lower ranking candidates after they’d been eliminated, Mills held onto her lead and won the nomination.
This brings us to the second takeaway from Maine, which is that RCV is not too complicated for voters to understand. If ranked choice voting were more confusing than plurality voting, we would expect the confusion to show up here, in a crowded, seven-person race, with a higher than normal number of “overvotes”—ballots that are filled out incorrectly and thus discounted. But there does not appear to have been any more overvotes in the gubernatorial primary than in previous elections, meaning most voters understood how the ballots worked. Consistent with that data, a Chamberlain Foundation exit poll found that four out of five surveyed were familiar with ranked choice voting before primary day; 69 percent said they found the process easy to use. Finally, the Maine Secretary of State’s office reported that turnout for the 2018 primary exceeded that of 2010, the last primary in which both parties had contested gubernatorial nominations. Even if RCV is more complicated than a plurality system, that complexity does not appear to deter people from voting.
What about the administrative concerns? Dunlap made a presentation to the National Association of Secretaries of State on Maine’s experience. He said that despite the rushed timetable and lack of resources, the overall process ran amazingly smoothly, though the final tabulations were delayed about a week while ballots were collected from around the state and paper ballots were integrated with those cast electronically. The increased scale did not dramatically decrease RCV’s efficiency.
When it comes to negative advertising and reduced campaign costs, however, the results are mixed. Candidate fundraising was fast and furious right up to primary day. In the gubernatorial primaries Cote raised and spent about $1 million; Mills and Moody each spent around $700,000, and Betsy Sweet spent more than that in public money. These are large totals for Maine races. In addition, EMILY’s List spent $300,000 in the last two weeks of the campaign, much of it in anti-Cote advertising aimed at helping Mills, whom they endorsed early in the race. In the Second Congressional District race, Golden raised over a million dollars and St. Clair about $700,000; St. Clair also benefited from independent expenditures supporting his candidacy. It is hard to argue that RCV reduced the costs of these campaigns.
However, there is some evidence that negative campaigning was less prevalent than normal. Of course, how one evaluates that depends on how one perceives the various campaign ads. Cote ran two ads that he characterized as “contrast” ads, noting differences between his positions and Mills’s record; she countered with an attack ad clearly aimed at him. More significantly, $200,000 was spent by an outside group, reportedly EMILY’s List, which specifically attacked Cote. But virtually all that money was spent in the waning days of the campaign, less a calculated political strategy than a last-minute Hail Mary. On the Republican side, candidates did go after Moody, but not with the rancor one often sees in campaigns like this one.
In the last weeks of the campaign Democrats Eves and Sweet campaigned together, asking voters to vote for them but give the other their second place vote. The Chamberlain Project poll reveals that one in five Democrats was asked by a candidate for a second choice vote if not a first choice vote. That is exactly the sort of positive campaigning RCV was designed to promote.
Were the gains worth the cost?
How one evaluates ranked choice voting depends on your goal. Republicans in Maine oppose RCV not because the system is flawed, but because they do not think that they can elect a true conservative candidate under that system. And they’re not entirely wrong. Ideologically pure candidates facing a group of more moderate candidates are more likely to win in a plurality system than an RCV system, although that’s true on the Left as well.
RCV in multi-member districts would benefit minority groups—racial, ideological, or otherwise—by assisting their path to representation in a legislature; in the same way, it would favor third, fourth, or other minor parties. RCV in single-winner races favors moderates over those at the ideological extremes. Whether one thinks this is a good or a bad thing depends on one’s political preferences; it’s a pragmatic question, not an ideological one.
As noted earlier, RCV in Maine will not be in effect for state elections in the fall; the procedure violates the Maine constitution’s procedures for elections for governor or state legislature. Two independents, one publicly funded and one with personal means to run an effective campaign, have qualified for the Maine gubernatorial ballot in November. The result in November is very likely to be a plurality winner with far fewer votes than needed for a majority. The same dissatisfaction that followed Governor LePage’s original election in 2010 is likely to reappear after this one—no matter who wins.
At the same time, however, Maine’s Federal elections—to Congress and the Senate—will be run under RCV (because the Maine constitution is silent on those races). Independent Senator Angus King is opposed by a Democrat and a Republican; First Congressional District Democratic Congresswoman Chellie Pingree is opposed by an independent as well as a Republican. Most observers predict that the two incumbents will win handily, but what is certain is that the voters will have a choice to cast a ballot without fear of wasting it.
What remains to be seen is how Maine’s citizens will respond to these very different experiences and whether they can mount a campaign to allow RCV for all elections. Then of course there’s the question of how the nation will respond to Maine’s example. For many years a common political adage held, “as Maine goes, so goes the nation,” because Maine voted early in presidential elections and usually ended up supporting the eventual winner. But in 1936, the adage was changed to “As Maine goes, so goes Vermont,” because Maine’s early vote for Alf Landon was copied only by its New England neighbor.
And for decades now Maine has eschewed a winner-take-all vote for the Electoral College, opting instead for a system in which the winner of each congressional district receives one vote, and the statewide winner two votes. Only Nebraska has adopted a similar plan—and so it may well be with ranked choice voting. David Brooks believes that RCV is the reform to save America. Maine has led the way in this experiment, ensuring that ranked choice voting will continue to be used there for some time. But still, the question remains: “As Maine goes,” will others follow?
1Maurice Duverger, Political Parties: Their Organization and Activities in the Modern State (Methuen and Co., 1964; original French edition, 1951).
2John F. Bibby and L. Sandy Maisel, Two Parties—or More? The American Party System, Second Edition (Westview Press, 2002).
3Thomas Hare, A treatise on election of representatives, parliamentary and municipal (1859).
4Committee for Ranked Choice Voting, “Maine’s History of Majority Rule and the Adoption of Plurality Provisions.” See also, Michael Shepard, “How an 1880 Maine Insurrection Could Sink Ranked-Choice Voting,” Bangor Daily News, January 21, 2016.