User:Lucasvb/Is Instant-Runoff Voting the right voting system for you?

Instant-Runoff Voting (IRV), which also wrongfully goes by the ambiguous umbrella-term "Ranked Choice Voting" (RCV), is the system currently at the forefront of the voting reform movement. Many people first get excited about voting reform through hearing about IRV, as I did, driven towards it due to several promises being made as to how it will address problems seen in politics and elections, problems which seem to plague all democracies all over the world today.

But before we get too excited, it is important to carefully understand what problems IRV is specifically designed to address, how it achieves that goal, and how it handles other problems.

This article is intended as a guide to those seeking to improve elections and politics by changing the voting system, but who don't know how they feel about the alternatives yet.

This guide assumes the reader already knows how the Instant-Runoff Voting algorithm works in detail. If you are not familiar enough, make sure to read the article on Instant-Runoff Voting first!

What do you want out of voting system reform today?

This is the question we all should ask ourselves. Please, take a minute to reflect. Think about everything you feel is wrong about politics today, things that ought to be fixed, things that ought to be thrown away, to be improved upon.

Ready?

Now Let's see how well Instant-Runoff Voting addresses your concerns on the following scale: BAD OK GOOD

Note: Many questions overlap, but their wording may reflect different approaches to the same issues.

I want to... Move away from the broken principles of FPTP and runoff elections

BAD

First Past the Post and runoff elections are both based on the "choose only one" principle: that the act of "voting" means the voter is naming only one candidate, and giving them , and only them, their full support. That is what causes most of the problems in both systems, such as similar candidates artificially weakening one another's performance via vote splitting, leading to things like the spoiler effect. In the case of runoff elections, candidates are also eliminated based on their amount of support, which means that candidates which are splitting votes are more likely to be eliminated, which is a similar spoiler effect on the first election round.

This "choose only one" principle typically forces voters to closely follow pre-election polls as to "not waste their votes". Since polls are conducted to reflect the results of the voting system that will be used, this also causes distorted pictures of "who's in the lead", "who's behind", "who has the most support", and so on. A candidate who hasn't split votes with others will tend be favored over other candidates that share a base and split votes, even if this shared base represents a large portion (even beyond a majority) of voters. This severely distorts the political narrative of which candidates and parties are viable and well-liked by the population.

IRV doubles down on every idea of FPTP and runoff elections, since every one of its elimination steps acts as a mini-FPTP election with elimination of the plurality loser. The elimination of candidates brings with it all issues with runoff elections on every elimination step of IRV.

IRV attempts to "fix" these problems by "unsplitting" the votes among the remaining candidates, but the damage is already done once any candidate is eliminated based on split votes.

So if you want to to move away from what broke FPTP and runoff elections in the first place, IRV is not the system for you.

I want to... Prevent smaller third parties from spoiling elections for the two big parties

GOOD

Due to the way IRV eliminates the "least favored" candidate in each elimination step, the process of "unsplitting" the votes tends to favor the larger or more mainstream party\candidate.

In this way, IRV will ensure the most favored big party will get legitimate support over the other big party, preventing a result in the opposite direction of what a majority wants, relative to the two big parties, that is. In a sense, IRV ensures the election is safe "for the big leagues" by preventing the smaller candidates from causing too much trouble.

This is its biggest benefit compared to FPTP today, as smaller parties may cause large disruptions due to the spoiler effect.

I want to... Get rid of the spoiler effect

BAD

Since candidates still split votes of their shared base in every elimination step, similar candidates penalize one another and risk being eliminated first. Isolated candidates still get an edge.

So under IRV, you still have a spoiler effect happening in each elimination step, it just looks different. The only difference is that you will be punishing the "smaller spoiler", not the "big spoiler", as it happens in FPTP and runoffs.

In other words, if you only want to prevent vote-splitting from "penalizing the mainstream", IRV is great. If you don't want anyone to be penalized by the spoiler effect, IRV is not the system you want.

I want to... Allow people to honestly express their opinion

GOOD

IRV's use of ranked ballots means people are allowed to be more expressive, offering much more information to the system than "name only one" voting systems like FPTP or runoffs.

However, this does not mean the information is actually used during the vote counting process. The information is in the ballots, but in a lot of ways it doesn't get used for anything.

I want to... Have the opinions of voters actually mean something

BAD

Since IRV only looks at the current top preference of voters in every elimination step, the vast majority of information in all the ballots will never be used for anything.

While each ballot is very expressive as a whole, in practice they always act exactly in the same way as "choose only one" ballots on every step.

So even though voters may have expressed more information, most of the opinions expressed typically do not matter and are effectively ignored.

In practice, this means IRV will simply and quietly remove smaller candidates from the picture every time, since any candidate's favorability as a second or third favorite counts for absolutely nothing in a given step, under the rules of IRV.

I want to... Allow people to express nuanced opinions about all candidates

BAD

The defining characteristic of ranked ballots is that they don't inherently convey nuance by themselves. Rated ballots, on the other hand, do allow for nuance to be expressed.

However, a ranked system which looks at the entire ranked ballot of voters in multiple ways may extract nuanced information out of a population. Condorcet methods do this, by making pairwise-comparisons between all possible candidates.

IRV, on the other hand, goes in the opposite direction. It explicitly looks only at the current favorite candidate, giving them 100% of a voter's undivided support, while completely ignoring everyone else.

I want to... Allow voters to always support their favorite safely

BAD

Due to still having vote splitting in every round and also being based on elimination of candidates, IRV will have many situations in which supporting your favorite is a bad idea for you and may lead to a worse situation. This will encourage voters to pay attention to polls and support the "lesser-evil"

I want to... Stop having voters supporting the "lesser-evil"

BAD

I want to... End concerns about "viability" of candidates

BAD

I want to... Allow voters to always support other candidates safely

OK

In Instant-Runoff Voting, you will never help other candidates unless your first choice gets eliminated. This means you can be sure that your current top choice will always get your full support on every elimination round, and therefore, you are free to rank the candidates in order of preference as they are eliminated.

However, this property also means you cannot always safely support your favorite first. There is a fundamental trade-off between "you can always support your favorite" and "you can always support candidates other than your favorite". The only voting systems that have both properties are exceptionally complicated, and not very good to begin with.

So you must choose whether always supporting your favorite is more important than always supporting other candidates.

I want to... Break free of polarization and partisanship by electing more consensus candidates

BAD

Due to vote splitting, IRV penalties consensus candidates and favors extremist, polarizing candidates.

I want to... Allow third parties to be taken a bit more seriously by the media

OK

Since the votes third parties receive must be transferred and this transfer is important to the results, this means third parties get more attention and participation in the media.

However, in places that have used IRV for a long time like in Australia, you still have political analysis and election results being presented as two-party preferred vote, which effectively ignores much of the information in the ballots and gives third-parties much less exposure. So in practice, it seems that IRV promotes a political bipolarization narrative in the media.

I want to... Allow third parties to be taken more seriously by the mainstream parties

BAD

Third parties are effectively rendered harmless, by design. For mainstream parties, any votes they lost just transfer back so it's in the interest of the mainstream parties to keep third parties small.

I want to... Get rid of in-fighting between groups who share a common cause

OK

The transfer of votes that happens under IRV tends to mostly reduce these fears.

However, since similar candidates still split votes and may risk eliminating one another, in highly-competitive scenarios this may still lead to some in-fighting and negative campaigning, which will act as a weak spot that can be exploited by opposing groups.

I want to... Allow third parties to thrive and be competitive

BAD

The way IRV is designed is to do just the opposite. Unless the third parties are somehow already more favored than a mainstream party, they will simply get eliminated every time. This imposes a very hard threshold for them to surpass to even get proper recognition.

This will also show up in pre-election polls, where third parties will get small percentage of supporters, reinforcing the idea that these parties are never going to make it big. This consistently reduces confidence of voters in supporting them, regardless or not if the voters will be transferred.

In effect, IRV was designed to deal with a "two and a half parties" scenario, and because of that it tends to perpetuate it. This is something that has been consistently observed in Australia, which has used IRV for over 100 years.

I want to... Have important consensus issues dominating the discussion

BAD

IRV favors polarization, as vote splitting still occurs. Whenever voters must take sides and the largest groups wins, the sides will have to maximally compete with one another even if they significantly agree. Overall, the result of this is that petty differences are highlighted much more than important issues and policy.

If a voting system is not designed to work within this "taking sides" paradigm, there is no such problem.

I want to... Have a more cooperative non-partisan political culture

BAD

IRV doubles down and amplifies the factionalism that we all hate today.

I want to... Have pre-election polls which reflect the honest opinion of voters

BAD

Pre election polls will be based on first preference + two party preferred votes, so the picture is heavily distorted.

I want to... Have pre-election polls which inform rather than disinform voters

BAD

The distortion discourages voters from promoting good candidates they believe in because they would rather promote another candidate with more chances first to prevent it from being eliminated.

I want to... Give voters the choice to support candidates weakly or strongly

BAD

Under IRV, candidates beyond your current top choice receive no support from you. Most ballot information doesn't even get considered in statistics due to this.

IRV is designed so that voters are always giving 100% of support to one candidate, and 0% to everyone else. Having multiple elimination steps doesn't change this.

I want to... Have election results which reflect the honest opinions of the population

BAD

Since only first preferences are important at any given elimination step, upcoming and rising parties may never be seen as such in the results. For example, let's suppose a small third party gets people excited and manages to secure 20% of the first-preference votes on the first step of elimination, compared to 30-40% of other mainstream candidates.

That upcoming candidate will be eliminated, and that will be the end of it. All the results will report the "20%" and nothing more.

But what if that candidate was the second preference of 80%-90% of the voters, reaching across the isle and gathering genuine bipartisan support? That information would never have been even talked about or seen by anyone, because preferences past the top choice are never looked upon without elimination of the top choice. Everyone who supported mainstream candidates only showed support for them and nobody else.

Wouldn't you agree that this would severely misrepresent the genuine opinion of the population about that upcoming party?

I want to... Have election results which represents everyone, not just a group of voters

BAD

Looking at only first preference is the same mistake FPTP is doing. IRV just extracts that information more accurately, but the mistake is the same.

I want to... Reduce the influence of campaign money

BAD

If polls didn't split the support of the population (they were cardinal) they would act as free advertising for good grassroots candidates. This will not be the case here.

I want to... Pave the way for better reforms in the future

BAD

Any voting reform that doesn't solve what it sets out to do wastes precious political capital and time. People will say "Why should we trust you voting system reformers if the last time it didn't work?" - "Oh, but this system is different! This one will work!" - "Yeah right, that's what they always say! Get lost, nerd. Leave your math out of my politics! "

I want to... Elect a candidate with majority of support

BAD

The majority of IRV is artificially constructed, because it changes the preferences of voters when it removes candidates. By the same logic of the system, you could elect the candidate with 100% consensus by eliminating everyone until one remains. This is clearly absurd.

I want to... Be politically progressive and bold and try something new in politics for a change

BAD

IRV is a super old system that has achieved very little in 100+ years of use. It is doubling down on every issue of FPTP, because the system is literally iterated FPTP with multiple elimination of the plurality loser.

I want to... Prevent strategic voting, where people try to impose themselves onto others

BAD

The way IRV addresses strategy is by making everyone as strategic as possible by default, maximally imposing their favorite on everyone in a stubborn way until it is removed by force.

I want to... Prevent strategic voters from getting an edge over "naive voters"

GOOD

Because IRV ensures there is no such thing as a "naive ballot". All voters are assumed to be stubborn maximally strategic voters.