By Kevin Oles
Instant Runoff Voting as a “Game Changing” Alternative Voting System?
Campaigns are exciting, emotionally charged political battles. Each year, as candidates fight for endorsements, contributions, and votes, it seems these political battles become increasingly bitter, partisan, and polarized. While “polarization” is a word with many uses in politics, I use polarization as a measure of the level of division of political opinion and rhetoric into, and along, party lines. Scholars who study this field report that polarization (both amongst candidates and voters) started increasing sharply in 1965, continuing to increase in every election cycle. Although individual candidates attempt to break this trend, campaigning positively, too often external forces pressure candidates into dividing the election and the electorate. Thus, if we think depolarization is a good thing, changing the systems that pressure candidates to polarize likely will decrease overall polarization.
Plurality voting, or “first past the post” voting is the current, familiar voting system used in most elections in the United States. Plurality voting systems define “winning” an election as receiving the most votes cast. In fact, some of the other election law commentators on this site have noted that plurality voting will be tested this year.
Instant Runoff Voting (IRV) defines winning differently, requiring that winning candidates receive a majority of votes cast. IRV requires that voters rank candidates: some IRV variations ask for voters’ top three candidates, while other IRV systems require voters rank the entire ballot. Votes are then counted in a series of rounds. In the first voting round, voters’ first place votes are counted; if a candidate receives a majority of votes, that candidate wins. If no majority arises, the candidate who receives the least number of votes is removed from the next round of counting. Voters who ranked the dropped candidate first have their votes counted in the second round based on their second choice votes. The second round of voting proceeds in the same manner: the votes are tallied and either one candidate receives a majority or the candidate receiving the fewest votes is removed (and votes are redistributed based on their next level of choices). This cycle repeats until one candidate receives a majority of votes.
While IRV seems convoluted, looking at an example of IRV voting makes the process clear. The 2009 mayoral race in Burlington, Vermont, was a three-candidate race (a Democrat, a Progressive, and a Republican), and after the first round of counting votes, no candidate received a majority:
Democrat Andy Montroll—29%
Progressive Bob Kiss—34%
Republican Kurt Wright—37%
The candidate with the fewest votes, Andy Montroll, was eliminated for the second round of voting; voters who ranked Montroll first had their second round votes counted based on their second choice votes. In the instant runoff second round, Progressive Bob Kiss received a majority of votes and won the election.
Looking at current three-way races, we begin to see the depolarizing effect of IRV. Using recent poll numbers, an IRV thought experiment demonstrates the difference in results between plurality and IRV. In Florida, if the Senate election occurred today under a plurality system, Republican Marco Rubin would win, but would not be elected by a majority of voters, merely the largest bloc of voters:
Marco Rubin (R)—40%
Kendrick Meek (D)—22%
Charlie Crist (I)—30%
IRV voting alters the results—no candidate receives a majority in the first round, so voting moves to a second round. Meek is removed and his voters’ second choice candidate becomes relevant.
Furthering this experiment, let us assume that voters who ranked Meek first would choose Crist as their second place vote. In the second round, head-to-head between Crist and Rubin, Crist receives an additional 22% of the votes (those voters who ranked Meek first and Crist second) and achieves a majority.
IRV
First Round Second Round
Rubin—40% 40%
Meek—22% (dropped) 0 (out after first round)
Crist—30% 52% (30 + 22)
The Florida thought experiment indicates the depolarizing pressure of IRV—its voting design has a direct effect on results. Consider campaign strategy under the plurality voting, where candidates attempt to do two things: (1) maximize the number of votes cast by their faction of voters; and (2) encourage non-faction voters to join their faction. Motivating voters through partisan attacks, negative campaigning, and polarizing the electorate are effective, proven methods of campaigning and winning an election. Candidates encourage their factions to think in an “us versus them” mentality that creates deep divides across partisan lines. Plurality voting simply does not provide candidates with incentives to bridge gaps, ally factions, or work across divergent opinions; rather, polarization increases unchecked.
IRV diminishes the value gained from encouraging voters to think in partisan categories, because the system requires voters to compare and contrast the entire ballot of candidates, thinking not in a binary “yes/no” manner. Thus, even though major party candidates could split the vote in a first ballot, an independent candidate could win in later rounds because IRV voters will prefer an independent candidate to an opposing party candidate with their second and third place votes. Thus, candidates under IRV do not achieve the same results when campaigning by polarizing the electorate as they do with plurality voting—polarization may ensure votes from their own faction, but the most divisive candidates will not earn necessary second and third choice votes.
IRV reduces gains from dividing the electorate because candidates cannot earn the valuable votes that come from outside their faction. Instead, IRV pressures candidates to appeal to diverse groups of voters instead of alienating them—decreasing partisanship through systemic change. Candidates in IRV cannot win by motivating their electoral base to voting en masse. Looking back to Florida’s Senate race illustrates this point: IRV allows voters to think of themselves in comparison to party and faction positions, and negative attacks have the distinct possibility of alienating voters. In fact, San Francisco’s IRV system has been singled out because of its ability to reduce partisanship between candidates.
IRV’s emphasis on uniting factions concords campaign action with our Founders’ beliefs about voting and factionalism in politics. Madison, in Federalist No. 10, wrote that while it was impossible to remove the causes of factionalism, a sufficiently large democracy would extinguish the kindled flames of any particular faction. Madison wrote that majority rule required that factions ally interests with other factions, mitigating the deleterious effects of a single faction. Madison realized that requiring a majority when voting makes it “less probable that a majority of the whole will have a common motive to invade the rights of other citizens. . . .“
IRV is not the panacea to all problems of polarization and negative campaigning. IRV’s depolarizing effect does not arise in head-to-head elections; a majority is assured in the first round, thus candidates can divide and polarize without the risk of losing the election. Additionally, IRV can also lead to a winner that is almost no one’s first choice. For example, as the second and third choice votes come into play, a candidate who squeaked into later rounds by being marginally better than the eliminated candidates can win by being every voters’ best alternative to giving votes to the opposing party. As a result, while a candidate might lose in a head-to-head race, under IRV, that same candidate wins as an acceptable alternative.
IRV’s has certainly become a “buzzword” amongst those who study elections, as Britain moves to IRV in 2011 and some U.S. cities have adopted IRV. It is clear that IRV can change current campaign tactics that result in bitter political fights and polarized elections. While immediately changing every election to IRV is not the answer, it is worthwhile to watch this developing voting trend and see what changes IRV does engender.