Representative democracy gives people the power to put their values into action. People can elect leaders who care about the things they care about and then hold them accountable for taking action on those issues. In the United States and Canada, people care about climate change. In Cascadia, people care about protecting their communities from dirty, outdated fossil fuels. So why are elected leaders not aligned with voters on climate change and other issues their constituents care about?

Because the way most North American governments elect executive officers—such as mayors, governors, and the president—doesn’t engage people, doesn’t empower people to vote their values, and doesn’t necessarily elect leaders with broad support. For example, only around 10 percent of US citizens chose Donald Trump and Hillary Clinton in the presidential primaries, leaving many general election voters feeling forced into choosing the “lesser of two evils” rather than participating wholeheartedly in electing a leader they could get behind. In part due to lack of good options on the ballot, ugly, negative campaigns, and the feeling that elected officials don’t represent them, the United States has dismal voter turnout rates.

Could a different way of voting give voters more voice, nurture more issues-driven campaigns, and elect leaders with broader appeal? Which voting systems would give Cascadian reformers better results? This article gives Sightline’s take on what is important in a voting system for electing an executive office held by a single person at a time (like a mayor or governor) and how different voting systems measure up. Although most Cascadian jurisdictions use the same voting method to elect the president and the legislature, the voting system options and considerations for electing legislatures are quite different because more than one person serves in the legislature at a time. (A future article will address voting systems for electing legislative bodies consisting of more than one person at a time—like congress, parliament, state legislatures, and city councils.)

What needs fixing?

The US and Canada use “vote for one” elections that limit voters’ choices and stifle healthy discussion of the issues people care about. Specifically, using the current system to elect mayors, governors, and the president, we suffer from the following problems:

  • Voters have to vote for the “lesser of two evils.” You should be able to vote for at least one candidate you support. Yet you must often vote for one of two front-runner candidates and not for a minor-party candidate, or a candidate who was eliminated in the primary, who you like better. You only have one vote, so you grudgingly give it to the least objectionable candidate you know has a shot at winning. When voters feel they don’t have a chance to vote for somebody they support, they may not vote at all. And, as we have seen, not voting can be as significant a move as casting a ballot.
  • Unpopular or extreme candidates can win. A candidate vying to be the only person representing a whole city, county, state, or country should have broad appeal. Yet, in the current system, a candidate who is unpopular with a broad share of the electorate can win. He can either split the majority vote and win with a mere plurality (in a three- or four-way race), or he can win over enough of the few, partisan voters who participate in party primaries to become the only viable option on the general ballot in a jurisdiction that is “safe” for his party.
  • Personal attacks work better than discussion of issues. Negative campaigning works in our “vote for one” system. Because voters can only express an opinion about one candidate, candidates are rewarded for turning voters off to opponents and aren’t rewarded for reaching out or building bridges to voters who have already chosen another favorite. Negative campaigning amongst a narrow field of candidates means voters don’t hear their issues discussed, and they may find little reason to tune into campaigns and engage in civic life.

Criteria for executive races

Political scientists and mathematicians have come up with many criteria by which to evaluate voting systems, resulting in complex tables like this one. But, unfortunatelyno system is perfect. Reformers have to decide what they believe is most important to fostering the healthiest democracy. (Note: The Institute for Democracy and Electoral Assistance has a tool that translates criteria into priorities and selects the best voting systems for you based on your stated priorities.)

We at Sightline came up with these four criteria to help determine how well a voting system can solve the problems described above:

  • Voters have a variety of options: Candidates with diverse views can run, and voters can vote for them without worrying they might be “throwing their vote away” on a minor-party candidate, say. Instead of limiting voters’ options, the election method tends to expand voters’ choices, welcoming more candidates and citizens to participate in elections
  • Winners are bridge-builders with broad appeal: A candidate with an energized but narrow base, unpopular with the broader public, can’t win. Instead, the election method tends to reward candidates who connect with a broad swath of the electorate
  • Campaigns are positive and inclusive: Candidates’ best strategy is to engage with many voters and discuss the issues that are important to them.
  • The change creates momentum for further voting reform efforts.

How each system measures up

Now that we know what principles rise to the top for Sightline, let’s look at how well different election methods do for our four criteria. (If you are unfamiliar with or need a refresher on different types of voting systems before you dive into the analysis below, see our Glossary of Voting Systems for Electing Executive Officers.) We have plenty of empirical evidence for how  Plurality, Top-Two Runoff, and Instant Runoff perform in elections. We have some evidence for how Approval Voting performs in student body and professional organization elections, but unfortunately, we have little or no evidence of how Score and Score Runoff perform, so we must speculate based on the systems’ characteristics.

(Click on the headings below to expand or collapse the discussion of each criterion.)

The most common voting systems in place in the US and Canada today are Plurality (vote for one, and the candidate with the most votes wins) and Top-Two Runoff (vote for one in the primary; the top two advance to the general; vote for one in the general election).

Additional candidates, beyond the two major-party mainstays, might appear on a Plurality ballot, but they usually receive very few votes because voters don’t want to throw away their only vote or spoil the election for their preferred major-party candidate. American voters all know how this works: if a majority of voters prefer both Nader and Gore to Bush, but split their votes between Nader and Gore, then Bush could win with a plurality (more votes than any other candidate, but less than half the votes). So voters cast their one vote for Gore, and Nader wins very few votes and often doesn’t even run.

Under Top-Two Runoff, voters see just two candidates on the general election ballot. The general election campaign is usually exclusively shaped by the two major parties, with no opportunity to bring other viewpoints into the discussion. Because they can’t make it to the general election, Top-Two Runoff discourages additional candidates from participating even in the primary.

Other voting systems allow more candidates to appear on the ballot and allow voters to express opinions about more than one candidate. However, only one makes it safe for voters to express an opinion about multiple candidates. Instant Runoff Voting (a single-winner form of Ranked-Choice Voting) lets voters rank candidates in order of preference on a single ballot and then simulates a series of runoffs. Under Instant Runoff Voting, it is safe to rank a weak third-party candidate like Nader (a third-party candidate who is in third or lower place). If you rank him first and he is eliminated, your vote transfers to your next-ranked candidate who is still in the running.

For example, say you ranked Terry Tea Party first, Larry Libertarian second, Ronald Republican third, and, just in case, Deborah Democrat fourth. Your vote would count for the Tea Party candidate in the first round; if she was eliminated, your vote would transfer to the Libertarian if he was still in the race; and if he was eliminated, your vote would transfer to the Republican if he was still in the race (probably, since he is a major-party candidate). If, by chance, the Republican had also been eliminated and the Democrat was running off against the Progressive, your vote would go to the Democrat. Or, if you ranked the Republican first, your vote would count for him in every round so long as he was not eliminated.

Original Sightline Institute graphic, available under our free use policy.

IRV can create a “center squeeze” situation where a candidate in the ideological center (likely a major-party candidate) could have won if a candidate to one side of him had not run, but because he ran, the candidate on the other side won. The one time this happened, in Burlington, Vermont, the Republican (on the right) and the Progressive (on the left) got the first and second most votes while the Democrat (in between them) in third place was eliminated. In the runoff between the Republican and Progressive, the Progressive won. But if the Progressive or the Republican had not made it to the runoff, the Democrat would have won against the remaining candidate. For example, if Republican voters who preferred the Democrat over the Progressive (about 17 percent of Burlington voters) could have somehow known that the Republican, who was polling in first place in a plurality election, would lose to the Progressive once the Democrat was eliminated, they could have strategically ranked the Democrat first to help him, not their favorite the Republican, make it to the runoff and beat the Progressive.

Because this situation rarely occurs (once in 142 US elections that release full data—0.7 percent of US elections—and no reports of it happening in a hundred years in Australia, though they don’t release full data so it is impossible to know for sure); and because it is difficult for voters to know ahead of time that their favorite, who is polling in first or second place, will lose; and because it is unlikely that voters will abandon their favorite and instead vote for a less-preferred candidate who is trailing in the polls; and because the candidate they would betray is a strong candidate—in first or second place—and the candidate who gets squeezed out is likely a major-party candidate; the center-squeeze situation is unlikely to discourage additional candidates from running in IRV elections or voters from voting for them.

The center candidate risks being squeezed out in any runoff system. If Burlington had used Top-Two Runoff, the Democrat would not have made it to the runoff. If it had used Score Runoff, he might not have made it to the runoff if Progressive or Republican voters had given him low scores (an average of 1.3 or lower) to protect their favorites.

Score Runoff Voting, Score Voting, and Approval Voting all let voters give each candidate a score. In Score and Score Runoff Voting, possible scores might range from 0 to 5, 0 to 9, or some other range, and all the scores are added or averaged. Under Approval Voting, the score is implicitly a 1 or a 0 because it is a vote or no vote; the ballot looks just like a Plurality ballot except you can vote for as many candidates as you want. Bucklin Voting uses a ranked-choice ballot but adds votes together like Approval Voting. All of these systems are flawed because they do not support what voting experts call “Later-No-Harm”: you can harm your favorite candidate by giving any other candidate a score or vote. When voters realize this, they often “bullet vote” (only score or vote for one candidate). If voters know which candidates are viable and which are not, they might vote for their favorite of the viable candidates and also any other candidates they like, so long as they are sure those other can’t beat their favorite.

Experience suggests that most voters using Approval and Score give their favorite candidate the maximum score or rank and all other candidates a minimal score or no vote. Candidates beyond the two major parties would likely not get many votes in these systems. Voters could figure out it is safe to give a minor-party candidate a vote or score as long as you are sure she will lose. But the major parties would encourage voters to bullet vote, and, to reassure voters it is safe to vote for them, minor-party candidates would have to convince voters they are sure to lose, which is a dog of a campaign strategy.

Score Runoff Voting should, in theory, encourage voters to give a maximum score to their favorite and also a score to their second-favorite, so that if their favorite is eliminated they could still get a vote for their second-favorite in the instant runoff. This would allow for a broader field as candidates ask voters for their maximum or a back-up score.

No single person can perfectly represent all the people in a city, state, or country. In other words, there is no perfect winner for an executive office held by just one person at a time. Different people have different ideas about who the “most right” winner is. The candidate whom a majority of voters support? The candidate whom most voters would choose over any other individual candidate in a head-to-head race? The candidate the fewest voters strongly object to (even if that also means that fewer voters strongly support him)? The candidate whom voters most strongly adore, even if many voters object?

No voting system can guarantee the winner is all of the above. For Sightline’s purposes, let’s set a lower bar: an unpopular or extreme candidate can’t win and become the only president, governor, or mayor for an entire country, state, or city. In this context, the definition of unpopular or extreme is: a candidate a majority of voters don’t want. One way of measuring this is that a majority of voters prefer every other candidate above this one (voting experts call this the Majority Loser criterion). Another measurement is that the candidate would lose to every other viable contender in a head-to-head contest, meaning a majority of voters would prefer any of the other candidates over him (voting experts call this the Condorcet Loser criterion).

If a voting system allows a candidate to win even if a majority of voters didn’t want him, it encourages candidates to fire up a narrow base of supporters while ignoring the majority of voters. This can lead to a divisive governing style and discontent with a system that would deliver such an unrepresentative executive. If a voting system guarantees that candidates without broad appeal cannot win, it encourages candidates to reach out more broadly to win over a majority of voters, as well as to govern more moderately, with the majority of voters in mind.

Instant Runoff Voting, Top-Two Runoff, and Score Runoff Voting meet the Majority Loser and Condorcet Loser criteria: they will never elect a candidate whom a majority of voters did not want or one who would lose to every other candidate in a head-to-head. The final head-to-head runoff in these three systems protects against an extreme candidate. However, in jurisdictions that hold party primaries and are “safe” for one party, an extreme candidate could win his party’s primary and then go on to win the Top-Two Runoff general election because voters are loathe to vote for the opposing major party. If instead Instant Runoff Voting or Score Runoff Voting were used in a single, high-turnout general election, it would elect a leader with broad appeal.

Plurality, Approval, and Score Voting fail both the Majority Loser and Condorcet Loser criteria: they can elect a candidate whom a majority of of voters did not want. In Plurality voting, this happens when the majority of voters split their votes between two similar candidates and the third, least popular candidate, wins with a mere plurality of the vote. The same can happen if most voters bullet vote in Approval Voting. Score Voting is, in a sense, designed to elect (this definition of) an unpopular or extreme candidate because it values intensity of preference over numbers of voters: a minority of voters can elect their favorite, even though he lacks broad appeal, by giving him a maximum score, beating the more broadly appealing candidate who received less than maximum scores from a majority of voters.

Bucklin passes Majority Loser but fails Condorcet Loser.

Negative campaigns don’t invite citizens to participate in civic discourse, and they take up airtime that could otherwise be spent discussing issues that are important to voters. Negative campaigns have other insidious negative effects, such as dissuading women from running for public office. To encourage positive campaigns, a voting system must reward candidates for reaching out to many voters, including those who might already prefer another candidate. Voting systems reward negative campaigns when a candidate can win by whipping up a narrow base of support, or can increase his chances of winning by insulting opponents.

Plurality Voting rewards negative campaigns because, to win, a candidate only needs to energize his or her base, not appeal broadly to a majority of voters.

  • Our work is made possible by the generosity of people like you!

    Thanks to Valerie Tarico & Brian Arbogast for supporting a sustainable Northwest.

  • Instant Runoff Voting has proven to produce more positive, civil campaigns. Because candidates can benefit from receiving second- or third-choice rankings, it is in their interest to court their opponents’ voters rather than ignore or alienate them. (Watch Minneapolis Mayor Betsy Hodges explain in this video.)

    Approval Voting could discourage negative campaigns as candidates urge voters to include them as one of several candidates they vote for. However, as discussed above, voters can harm their favorite by voting for additional candidates, leading to the real-world experience that Approval Voting ends up looking a lot like Plurality Voting. So candidates might campaign just like Plurality.

    Score Voting could lead to even more negative campaigns than Plurality Voting. Like Plurality Voting, Score Voting rewards candidates for energizing their base, but it additionally rewards candidates for provoking lower scores for their opponents. A candidate could win a campaign under Score Voting by whipping her the base into a frenzy of maximum scores while sowing doubts amongst her opponents’ supporters. Rather than moderating her message to win scores from more voters but risking alienating her base, she can focus on tearing down her opponent to both fire up her base and undermine other voters’ confidence in their candidate enough that they give that other candidate less than the maximum score. Score Voting could prompt a race to the bottom in terms of negative campaigning.

    Score Runoff Voting, a hybrid of Score Voting and Instant Runoff Voting, would likely be somewhere between the two in terms of campaign tone. Because voters would want to give some non-zero score to candidates other than their favorite, it would be in a candidate’s interest not to alienate her opponent’s voters in hopes they will give her a score. But as in Score Voting, a candidate could be harmed if her voters give another candidate a score, so she would want to encourage them to give her the maximum and give her opponents a minimum score, creating an incentive for painting her opponent in a negative light.

    Any reform must overcome the inertia of the status quo. Any reform that gives minor parties more power risks pushback from the two major parties. Facing into these headwinds, reformers need to know that each hard-fought win will build momentum for the next.

    In the United States, Ranked-Choice Voting is the only voting reform with momentum. Thirteen US cities and counties already use it; the state of Maine just adopted it for state and federal elections; and here in Cascadia, Benton County passed an Instant Runoff Voting initiative in 2016. In 2017, 18 states have introduced ranked-choice voting bills—11 states have Republican co-sponsors, and 13 states have Democratic co-sponsors.

    Instant Runoff Voting (a.k.a. single-winner Ranked-Choice Voting) can also build momentum towards a more powerful reform—proportional representation in multi-winner elections. As more American voters become familiar with a ranked-choice ballot, it could make it easier to introduce the system into a multi-winner ranked-choice election (a.k.a. “Single Transferable Vote”)—more on this in our Guide to Voting Systems for Electing Legislative Bodies, publishing next week.

    Score Runoff Voting is building momentum in Oregon. It also has a multi-winner form, but it may not achieve as fair results as multi-winner ranked-choice voting—more on this in our Guide to Methods for Electing Legislative Bodies,.

    Conclusion

    Based on what’s currently failing or out of balance in executive elections in the United States and Canada, Sightline judges Instant Runoff Voting as currently the best reform to pursue. It allows a diversity of candidates from major and minor parties to run and win votes. It nurtures more positive campaigns. It rewards candidates for building bridges to many voters and blocks extreme candidates from becoming the only mayor or president that a city or country has. And it has momentum in the United States.

    Score Runoff Voting deserves experimentation to see how it performs in the real world.

    Top-Two Runoff and Approval Voting offer improvements over Plurality Voting, but not as much improvement as Instant Runoff and Score Runoff.

    Plurality Voting is clearly an inferior option on every criteria. Score Voting also seems particularly undesirable, because it might incentivize even more negative campaigns and result in divisive, extreme candidates winning office even more often than does Plurality Voting.

    Our current “choose one” method of electing mayors, governors, and presidents is fraught with problems. But Ranked-Choice Voting, and possible Score Runoff, could give voters more options on the ballot, elect executives with broad appeal, and generate more positive, issue-oriented campaigns.

     

    A note about partisan gridlock, sound policy-making, and reflective representation

    Astute readers may note that the criteria above do not address the pressing problems of overcoming partisan polarization and gridlock, electing more diverse representatives, and fostering problem-solving and sound policy-making. Sightline passionately believes in reforming voting systems to solve these problems. They all fall within the provenance of the legislative branch, and we discuss them in our forthcoming Guide to Methods for Electing Legislative Bodies for electing multi-seat legislative bodies such as federal, state, and provincial legislatures and city and county councils. Better voting systems for electing legislative bodies can elect diverse bodies that reflect all voters, incentivize consensual problem-solving over partisan gridlock, deliver broadly supported policy solutions, and help strengthen civic life.

    Download a printable version of this guide