This blog post was originally posted to the MIT Election Data and Science Lab (MEDSL) Medium site, and may be found here.
On June 12th, Maine will be the first state to vote entirely using ranked choice voting (RCV). While this voting method dates back several decades and is used in 11 municipalities of various sizes around the country, it has never been implemented at the state level. This has caused a heated debate in Maine politics around the “true” aggregation of preferences and the legality of the voting method, and has been the subject of several lawsuits that have just recently been settled. In this post, I provide a broad overview of the debates around RCV in Maine, as well as some of the academic literature behind RCV in other contexts and how it applies to this case. In doing so, I seek to demonstrate that the implementation of RCV in Maine presents new and yet unknown challenges for campaigns and researchers. There are many questions left unanswered about how well the new system will work in Maine and in other contexts.
In 2016, a citizen initiative known as “Maine Ranked Choice Voting Initiative,” or Maine Question 5, was approved by 52% of voters. Like other RCV plans around the country, this system requires that voters rank the candidates for each office, from first to last. Counting is done by first tallying up the number of #1 ballots cast for each candidate. If a candidate achieves a majority of #1 votes, then they win the election. If not, the candidate who received the fewest #1 votes is eliminated. Ballots that ranked the eliminated candidate as their first choice are added to the number of #1 ballots for the second ranked choice for that particular voter. The first-place votes are then re-tallied to determine if one of the candidates has broken the 50% threshold. If not, this process of elimination continues until one candidate receives a majority of the #1 votes. More information on ballot counting and marking (including spoiled ballots and under-voting) is available from the Maine Secretary of State here and here.
Arguments in favor of RCV
Over the past two years, the League of Women Voters and the Committee for Ranked Choice Voting have run campaigns to pass and preserve ranked choice voting in Maine. In doing so, these groups have used several arguments in favor of the new system. First, proponents claim that it allows a majority of the electorate to choose the winner of an election, which they say is a key part of America democracy. They note that 9 of the past 11 gubernatorial races in Maine were decided by a plurality, and that there has not been a governor elected to their first term by a majority in four decades. Second, they claim that it will lead to more civil campaigning. Proponents claim that this is due to the fact that candidates will be encouraged to reach out to a larger swath of voters to earn more second and third-choice votes.
A third argument is that RCV allows voters to choose candidates “based on ideas, not on the most well-funded candidate.” This is the most heavily argued point by proponents of RCV, who seek to prevent the “spoiler problem” where a majority bloc of voters is split by two candidates of similar ideology, allowing a candidate from the opposite ideology to win through a plurality. As political scientist Jack Santucci notes, one of the major reasons for the push to end the spoiler problem comes from the low popularity of Maine Governor Paul LePage, who was able to win the 2010 Republican primary with 37% of the vote, and the general election with 38% of the vote.
Arguments against RCV
While proponents of RCV came out on top during the 2016 referendum, opponents of the new law have waged a legal and public relations campaign that has relied on several key arguments. Opposition to the measure has largely emerged from party structures, particularly the Maine Republican Party and the Senate Republican Caucus. This is unlike the proponents of RCV, who have primarily come from civil society. Maine’s RCV opponents focus on three arguments:
that the new system will be more expensive to administer than the traditional system;
that RCV will suppress votes, both from reduced turnout and through votes discarded in successive rounds of counting; and finally,
that it violates the Maine Constitution, which they claim requires elections to be decided by a majority or plurality.
Interestingly, the party’s official stance is that RCV will not change the results of elections or provide advantage to one side, in stark contrast to the “spoiler problem” argument of RCV proponents. Beyond the state Republican party, several candidates for governor have also argued against RCV. Gubernatorial candidate Mary Mayhew and several others, for example, argue that the new system will be confusing to both voters and election officials, and will likely not be implemented correctly.
In addition to a fairly decentralized public relations campaign, opponents of RCV have brought legal challenges against its use in the June 12th primary and beyond. These have come from both the Maine Republican Party in Federal court, as well as the Maine State Senate through the Senate Republican Caucus in state court. The lawsuits alleged that RCV will change the very nature of the Republican Party itself. It will alter the message of the party’s nominees, and thus strike “at the heart of the party’s right to freedom of association.” However, the Maine Supreme Judicial Court passed down a unanimous ruling in April against the suit from the Senate Republicans, and a Federal Judge ruled against the state party in May, allowing RCV to be used in the upcoming primary.
What do political scientists have to say?
While there has been a strong debate around RCV in Maine and other contexts, political science has had much less to say about ranked choice voting in the United States. According to Jack Santucci, party cohesion in multi-winner elections (such as in the Cambridge, MA city council) can be affected by RCV, depending on the outcome of an election. In cases where there is a large vote difference between candidates within the same party, cohesion will be high, while cohesion within a given party will be low in more competitive elections. In a recent working paper, Santucci leverages the 2016 RCV referendum to explain the conditions underwhich a political actor will support RCV. He finds that a party will support RCV when it is losing elections, when a “spoiler” is present, or when a party believes that RCV will lead to victory through majoritarian outcomes.
In terms of campaign civility, Sarah John and Andrew Douglas recently found that voters surveyed in cities that use RCV perceived that campaign tactics were less hostile, though it was not clear whether this was caused by RCV or is simply an artifact of the single-party politics in those cities. Furthermore, John and Douglas found that voters who voted using RCV tended to favor using the system more than voters who voted using traditional ballots.
Lindsay Nielson also recently published a survey experiment on RCV, which provides us with results that are more useful in analyzing the upcoming RCV election in Maine. She provides three major takeaways. First, she found that RCV had no measurable impact on voter confidence, which is the belief that the votes in an election were counted as the voters intended. Second, there was no difference in the outcome of the simulated election between the treatment and control groups, suggesting that there is little difference in the outcome between RCV and plurality elections. Finally, most voters, when put in a hypothetical situation using RCV, did not think that RCV elections resulted in fair elections, and those who voted in them were no more likely to support RCV than majoritarian elections.
How is June 12th going to play out?
Simply put, there is no clear indication how Maine’s June 12th primary will play out, in terms of which candidates will win. However, we can piece together enough evidence from polls, election administrators, and academic research on previous electoral reforms to come up with some predictions about how voters will handle the new system, as well as how the Secretary of State’s office will handle the new system.
First, the final results for the high-profile races are unlikely to be released until a day or two after the election. There are several reasons for this. Chiefly, recent polls have shown that there are no candidates in either the Republican or Democratic gubernatorial primaries receiving more than 50% support, nor are there any candidates polling above 50% in the Democratic primary for the second congressional district. If these results are reflected in the final vote count, then successive rounds of counting will be carried out by the Secretary of State’s office in Augusta. According to the Secretary of state, this will require all ballots to be transported to Augusta to be tabulated. This is no small feat, as some ballots may take over five hours to arrive in Augusta once the decision is made. For example, if it takes two hours to count the votes and decide to conduct the ranked count in Augusta, the final tabulation would not be able to begin until at 3am on June 13th, at the earliest.
Second, there will likely be a higher residual vote rate due to spoiled ballots and ballots that are not counted due in some rounds to lack of rank. While it may be a stretch to compare the implementation of RCV in Maine to ballot reform and vote by mail in California, the underlying link between ballot reform and residual vote rate appears to be fairly strong. As seen in Alvarez, Beckett and Stewart in the case of vote by mail in California, the number of spoiled votes and the residual vote rate increased, likely due to the confusing nature of the ballots. As such, we should expect confusion around ballot structure in Maine to lead to an increase in spoiled ballots and ballots not counted in different rounds of ranked tabulation.
Finally, there will not be an increase in voter confidence. If anything, voter confidence will be reduced. This is consistent with the findings in Alvarez, Beckett and Stewart, which demonstrated a significant decrease in voter confidence as a result of ballot reform, as well as Nielson’s research cited above, which found that RCV had no effect on voter confidence. A likely scenario is that the “winner’s effect” is prevalent, with voters who ranked the winning candidate as their #1 choice expressing more confidence in the results, and those who ranked the winner lower on their ballot expressing less confidence.
Overall, RCV will likely continue to be the subject of spirited debate in the state of Maine beyond this primary season. While I have attempted to use contemporary research to predict what will happen on June 12th, political science does not have a strong enough grasp on ranked choice voting to generate many strong predictions. While this may be a current gap in the field of election science, the case of RCV in Maine provides a rare opportunity to learn more about voting systems and behavior from a variety of angles, such as voter confidence, election administration, residual vote rate, and campaign strategy. If rigorously examined, researchers may be able to better inform the debate around ranked-choice voting in future contexts.