Why We Need Range Voting and Not Ranked Voting

Since I first heard of Instant-Runoff Voting (IRV), I’ve been a proponent of alternative voting systems. We’ve all seen first hand this election how having one and only one vote can mean the choice between the Giant Douche and the Turd Sandwich. To fix this problem with our current electoral process many different voting schemes have been proposed, the most popular and recognized of which is IRV, which recently passed in the state of Maine (and will be used in all of their elections, minus the president).

However IRV is not all it’s cracked up to be. It fails an important criteria of a proper voting system: Independence of Irrelevant Alternatives (IIA). The meaning of this is simple: say you have an election between three or more candidates. Removing any of the losers should not change the outcome of the election. For example, the 2016 presidential election was between Donald Trump and Hillary Clinton, however Jill Stein was also on many ballots. To meet the IIA criteria, the removal of Jill Stein from the ballot should have had no effect.

The importance of this is that voting systems that meet IIA are immune to the ‘spoiler effect’, which third party candidates are often accused of causing. That is, third party candidates can, er… ‘steal’ votes that would have otherwise gone to major party candidates. This is obviously very bad. Voting for your favorite choice should never result in a less favorable outcome. Even though the major party candidates never really ‘owned’ your vote in the first place, the reality is by virtue of voting for your first choice over your second choice you risk letting your least favorite candidate win.

Instant Runoff Voting is supposed to alleviate this issue. The concept of IRV, or ‘ranked voting’, is simple enough: You rank your candidates, and if your top-ranked candidate loses, your vote ‘instantly runs off’ to your next choice, and so on. An IRV ballot may look like this:

 Hillary Clinton2Donald Trump1Gary Johnson3Jill Stein

On this ballot I’ve ranked Gary Johnson as my number one choice. This means Gary gets my single vote first. Gary lost so my vote would then instantly run off to Donald Trump. He won, so the runoff of my vote would end there; Jill Stein would not receive any vote from me. If Donald Trump lost, Jill would indeed get my vote, but under no circumstances would Clinton get my vote, as I haven’t ranked her on my ballot.

So at first glance, it seems great. I can vote Johnson without allowing my preferred major party candidate, Donald Trump, to lose, as if my favorite candidates, Gary Johnson, loses, than my vote flows on to Trump. This allows me to hedge my bets between my two  or more top choices. Despite being better than our current system, IRV still fails to meet the IIA criteria. It’s all fun and games when you have a candidate like Gary Johnson which draws support from both major parties with rough parity.

However, what if there were a candidate which drew from mainly one party, like Dr Jill Stein of the green party? Most of her voters would rank their ballots with Jill as number one and Hillary as number two. That is all fine when Jill has no chance of winning, but what if she does?

That would mean rather than Jill’s votes running off to Clinton, it would be Clinton’s votes running off. However, there is no guarantee that Clinton voters would rank Jill Stein as number two. They may have ranked Gary Johnson as number two, or they may have no candidate ranked second at all – either would allow the other major party to obtain an easy victory. In fact, it would only take a small percentage of Clinton voters not voting for Stein as their second choice to prevent either Clinton or stein for winning.

Thus, IRV does not remove the spoiler effect and thus does not satisfy the IIA criteria.

However there is a better way: Range Voting

Range voting is where you give each candidate an approval rating, say, from one to nine. The ratings are added up, and the candidate with the highest rating wins. Here is an example of a range voting ballot:

Hillary Clinton0123456789Donald Trump0123456789Gary Johnson0123456789Jill Stein0123456789

As you can see, range voting is by a wide margin superior at representing my feelings about the election. I can hedge my bets by voting for multiple candidates, without fear of any spoiler effect. Furthermore, it isn’t a binary scale. This allows for quick, easy candidate ranking without having to choose which candidate is the best, which is the next best, etc. The non-binary scale also allows you to hedge your bets with a candidate you only marginally approve of without having the chance of giving them a full vote.

Range voting fulfills the IIA criteria: Voting a high score for your favorite candidate can never get you a worse outcome, unlike with IRV, where a Green Party voter could theoretically end up with the red candidate  by putting the green candidate before the blue candidate.

With this method, our green party voter could simply tick Jill Stein and Hillary Clinton with a 9/9, or they could put Jill with a 9 and Hillary with an 8, or any variation that satisfied their feelings of the candidates. As such, range choice voting is the ultimate ‘hedge your bets’ voting. You can never go wrong by giving a candidate you like a high score, as that only increases their total score and thus their chances of being elected.

This is especially important if there are two candidates that you like equally. With IRV, you’d have to put one or the other in front. With range choice, you can give both the same score. This works to increase voter expression.

Conclusion


IRV is not a bad development on our current system, but it is clear to me that Range Choice voting is definitely the way to go. Implementing alternative voting systems will go a long way in making our elections fairer and less biased against challenger candidates, and less biased for establishment candidates.

4 thoughts on “Why We Need Range Voting and Not Ranked Voting

  1. In order for Range to pass IIA, your ballot would have to be the same if Hillary weren’t on the ballot. I really doubt you would have given Stein a 7 if it were just the other 3 candidates, would you? So Hillary’s presence made you reduce the daylight between Johnson and Stein to a mere 1/3 of your vote’s strength.

    This does not seem particularly independent to me.

    You haven’t given consideration (at least, in this article – tbf, I haven’t read the rest of the blog, though given its name I certainly shall) to those ranked systems actually optimized for eliminating IIA, those which treat the races as independently as possible within the constraints of Arrow’s theorem. That is to say, Condorcet methods: Schulze or Ranked Pairs (or if simplicity is most important criterion, the nearly-equivalent-to-either-since-they’re-so-similar Smith-Minimax).

    Like

    • Thank you for commenting

      I really doubt you would have given Stein a 7 if it were just the other 3 candidates, would you?

      I gave her a 5? Also, yes. If Hillary was removed from the ballot maybe I wouldn’t feel the need to put Stein at 5, and give her a zero instead – but that’s only because I never really wanted stein in the first place, just didn’t want Hillary.

      As for the other methods, they were beyond the scope of the article. Thank you for bringing them to my attention however.

      Like

  2. but that’s only because I never really wanted stein in the first place

    Exactly – your ballot itself – and thus, obviously, its effect on the other races – was impacted by Hillary’s presence. Range may be IIA in respect to fixed ballots, but ballots do not occur in a vacuum. Ratings change based on context, and the kinds of transformations in IIA are the kinds of transforms that will change Ratings.

    As for the rest, the title is ‘Why we need Range voting and not Ranked voting’. Those are ranked systems. If the title was ‘Why we need Range voting (or something) better than Instant Runoff voting’, then sure, those would be out of scope.

    Like

    • Exactly – your ballot itself – and thus, obviously, its effect on the other races – was impacted by Hillary’s presence.

      Since responding to your comment, I have since realized you are right.

      However, I stand by my original point that range voting is superior to IRV for the stated reasons.

      As for the rest, the title is ‘Why we need Range voting and not Ranked voting’. Those are ranked systems.

      I don’t understand this at all. If you’re trying to pull a definition switch on me you very well know that by writing ‘Ranked Voting’ means IRV. Are there other ranked systems? Sure, but sometimes you have to sacrifice accuracy to get the point across in a headline. Furthermore, I use IRV by name throughout the entire article, so I don’t see how you could possibly say that this is in scope.

      Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s