We don’t need a third party. We need ranked choice voting
chicagotribune.com
chicagotribune.com
Ranked choice alone makes multiple parties viable but it's not ideal. Yet ranked choice would almost certainly lead to more than two parties in the US.
Electoral system is not only the voting system. Single member districts + plurality (current US system) is the worst system for representative democracy by multiple criteria, so almost any proposal improves the situation.
Proportional representation (PR) is any system where seats are divided between parties/lists by their vote share. Ranked choice is not enough. The US would also need multi member districts.
>By federal law, all members of the United States Congress are elected from single-member districts, and most states also elect their state legislators from single-member districts. Some states, however, utilize multi-member districts. For example, Arizona, New Jersey, South Dakota, and Washington elect all members of their lower state legislative chambers from multi-member districts. https://ballotpedia.org/Multi-member_district
Partly that's for good reason, since the public has seen how those aspects of the electoral process have been manipulated for partisan benefit in the past (or at least, they've been told that the other side is guilty of that, and they believe it) and it's partly because it's easy to argue against any changes by using FUD techniques, relying on people's mindset of "if it ain't broke".
So it's probably right to focus on a change that's limited to just the balloting and winner-determining processes for now, hoping that a future generation of more engaged voters and more representative politicians can push for changes that lead to more proportional outcomes, as you say.
The corollary of this logic, though, is that it would be better to change just the winner-determining process and not change the ballots at all. While I do think that RCV is a huge improvement to FPTP in the US, I do have some sympathy for the claim that RCV is more complicated, which can have a disenfranchising effect.
For example, it introduces unfamiliar new tactical dimensions, where people might be encouraged to rank their least preferred candidate disingenuously higher than another rival in order to cause that rival to be eliminated first. More generally, it pressures voters to have to produce an exhaustive ranking of all possible candidates, some of whom they might not have heard of, and potentially penalizing voters that opt to leave some ranks unspecified. Then there's the problem of needing to use untrustworthy electronic voting machines (or a much longer manual process) to count ballot papers that have N-factorial possible markings for N candidates, in a process which strongly encourages centralization rather than being distributed.
The solution to these problems is a method called Asset Voting, where ballots work exactly the same way as under FPTP, but at the end of the count, the candidates get eliminated one by one and can re-assign their votes to one of the remaining candidates. This completely removes the spoiler effect and the problem of "wasted" votes, while not introducing any extra cost or delay (other than two minutes of asking the eliminated candidates who their votes should be re-assigned to). Unfortunately, though, I don't think this method has the same momentum behind it as RCV or other popular alternatives like Approval Voting.
Most of these systems can be made better with multi-member districts. At this point D'Hondt-method is tried an true. The US could use Jefferson method (invented Thomas Jefferson) because "Murica" but it gives the same result :)
I agree with you, that the most important thing is to implement some change, not strive to fix everything at once. Once voting methods change, politics change and other changes can become easier.
I'm not sure what guarantees you require a voting system to provide. Do you not agree it is better than plurality voting?
I suppose you could say that it's not "guaranteed" to solve the spoiler problem, because if you vote for a candidate who claims to have very similar policies to another candidate you like, they might turn around and betray you and re-assign your vote to another candidate you are completely opposed to. However, the fact that candidates might talk one way but act another way seems to be an issue that's common to all voting systems, and therefore out of scope.
> At this point D'Hondt-method is tried an true.
Any electoral change that requires an explanation of the form "First we calculate the quotient with the following formula..." is doomed to failure in a world of misinformation and vested interests trying to undermine confidence in elections. You're right, though, that in the US it would be wise to advocate for Jefferson's method, and brand it that way.
You're probably right, though, that candidates should be forced to do this, and maybe they should be able to choose whether or not to disclose their preference in advance to potential voters. If they opt to keep it secret and they win, then their preference never has to be revealed, and if they are eliminated then only their highest remaining preference is revealed.
Admittedly I haven't proved that there aren't times when a candidate may need to tactically reorder their preferences based on the actual distribution of votes, but I can't think of a situation where they wouldn't want to re-assign their votes to the candidate whose values are most similar to theirs / their voters.
I think a possible rule to defend against it would be to mandate that any political party needs to have received at least 5% of the popular vote in that local region in the previous local elections in order to be eligible to stand candidates for the national congress/parliament.
If a candidate wants to stand without the backing of any political party, they would need to receive at least 20% of the (first round) votes in their district in order to win (without needing to stand in the local elections first).
Forcing candidates to publish their vote assignments at registration time might be the way to go, though, as it would help to expose these straw candidates, and also mean that the candidates wouldn't have to be present when the voting result is announced.
To avoid embarrassment, candidates could place a "null" option in their lists, meaning "If I am eliminated, then discard all the votes I received (including the votes assigned to me from earlier-eliminated candidates) and let the remaining candidates win or lose based on the remaining votes".
Just as´ examples of desirable properties are Unanimity, Independence of Irrelevant Alternatives. If you know enough to suggest asset voting, I suspected you know enough about social choose theory or electoral systems to have at least some of them in mind. Asset voting is just plurality voting that reverts into weighted electors. It allows any kind of surprises that voters did not want.
>Any electoral change that requires an explanation of the form
The simplest explanation for is
1. Each party gets representatives relative to the number of votes they received.
2. Party representatives are picked in the order they received votes (open list).
small print: how it is done mathematically was explained by Thomas Jefferson, in a letter to George Washington in 1792. Belgian mathematician Victor D'Hondt found simpler method in 1878. The methods are equivalent.
I'm aware of many desirable properties, such as the ones listed here[0], but "Unanimity" is not mentioned on that page. I struggle to even imagine what such a criterion would be. "If every voter votes for a given candidate, then that candidate should win"? That sounds like an obviously desirable property, but of course Asset Voting has that property, so I'm still not sure what the objection is.
The other property you mention (IIA) is one I'm not actually convinced is desirable. It makes the analysis of voting systems easier if you can assume it as an axiom, but I think it only applies to spherical voters in a vacuum. People change their preferences due to minor "unrelated" pieces of information all the time, so it doesn't seem odd to me that a voter might see party C on a list and update their view of the world, such that they decide they were wrong to support party A over party B, for example.
Anyway, you said "Asset voting gives no guarantees whatsoever", which I guess must be hyperbole, because Asset Voting is guaranteed to provide the desirable property of the Majority Criterion (MC), since a candidate with a majority of the (first round) vote cannot be beaten no matter the distribution of votes for other candidates.
The best I can do at steelmanning your complaint is that some of these desirable properties are difficult to reason about in relation to Asset Voting, because its result-generating process takes as part of its input the decisions made by the candidates themselves, which these formalized properties don't consider. However, that seems like a failure of the properties themselves rather than the voting method, unless you can think of a way that Asset Voting fails in practice due to their inapplicability.
[0] https://en.wikipedia.org/wiki/Comparison_of_electoral_system...
- 'new political party shall become an "established political party"'
and
- Sec. 7-19. 'The primary ballot ..' shall be arranged and printed ...:
1. Designating words. At the top of the ballot "REPUBLICAN PRIMARY BALLOT"; "DEMOCRATIC PRIMARY BALLOT"'
It is State election law which creates the domination of the two-party major/minor system. It's simply hard-coded this way. Coders might say RTFM.
It might be possible, for instance, to start a third party that exists solely to pressure the Democrats into supporting RCV (first at the state level, then federally).
And of course one of the main benefits of RCV is that it can allow a third (and fourth...) party to gain traction, although it can also improve politics just by making the existing parties more representative of their voters, since they won't be able to rely on the "lesser of two evils" effect.
Yes, and the more votes that third party gets, the less likely it is that the democrats have any power and therefore the less likely that they fulfill their mission. Which is the whole point why you need RCV in a system like the American. There is no way to move gracefully towards an improvement to the system, because your only options work counter to your goal right until the point where you have a full and sole majority in which case you didn’t change the two party problem, you just replaced one of the parties.
"Ranked choice" has a nice ring to it when giving someone the elevator pitch, and the name implies the voter is given more authority which is true. I find this easier than trying to start by explaining an acronym. Once they get engaged then i pivot to other methods.
I think that a lot of advocates for reform (especially on a site like this) tend to think of voting as a purely abstract exercise, and that the ideal reform is one which meets various mathematical criteria, but that seems to be the same sort of mistake as thinking that the best software is the one which uses the most efficient algorithms.
STAR voting, for example, may be more expressive, in that it allows the voter to input more information into the process, but I'm not actually sure that's a necessary goal, and it may even be an anti-goal, since the more input is required from a user, the higher the cognitive load, the more scope for things going wrong, the more attack surface for someone to spread misinformation, and the more error-prone / time-consuming / expensive the voting and counting process is.
Given that a vote typically has thousands of bits of information in aggregate, even with FPTP ballots, and the output only needs to be a single winner, it seems that allowing more input isn't actually a benefit. Instead, what is needed is more honest input, i.e. resistance to tactical voting (which I concede STAR voting is probably better at than RCV).
Unless we want to make elections forever dependent on voting machines whose firmware is written by the lowest bidder (or biggest campaign contributor) we should always prefer voting systems where the ballots can be counted by hand, in as distributed a way as possible. A team counting plurality ballots can stack them into N piles, for N candidates, and the heights of those piles are roughly proportional to the votes for each candidate. To do the same for RCV requires N! piles, and for STAR voting requires 6N piles. Approval voting requires 2^N piles (so it is preferable to STAR if there are 4 or fewer candidates) whereas asset voting requires just N, like FPTP.
Obviously in the US you can't do that in speech, so candidates can still publicly align themselves with others or on positions. But when you go to the ballot you should know who you're voting for and not base it on the D or R next to their name. Remove party affiliation from ballots and stop supporting the primary system and that would be an enormous improvement.
No wonder countries won't try it, party affiliation is their most effective means of controlling the narrative that keeps them in power.
i have been making similar arguments here before: https://news.ycombinator.com/item?id=25031034
but i'd like to go even further and eventually eliminate candidates as well: https://news.ycombinator.com/item?id=32267723
What the politician says or thinks doesn't matter at all. It only matters what the leadership of the party they are member thinks.
Political parties are creatures in the election laws of each State. The sections of State law which create (and then regulate) parties can be removed.
You can possibly remove laws which create specific parties but you can't make creating political parties altogether illegal.
This doesn't run afoul of their free speech, Candidate A can still say they are buddies with Candidates B,C & D etc. But when people go to the polls, they shouldn't rely on gang colors to determine who to vote for. Names only.
State election law in Illinois and others disagree with you. I already posted Illinois.
Look in your State laws for major/minor party, and they often set some % for the Governor race. I don't like the way it is coded, but that is the current law for the States I have researched.
Political parties get special benefits which encompass keeping down the third parties or minor parties. It's not like a group of randos who play cards and got a bank account to write a check to rent the bingo hall - there are special benefits to the major political parties in the election laws.
That's a good point about Illinois election law. I suppose if I were king for a day I would add an amendment to the US Constitution that specifically prevents this.
They shouldn't, but they have an obvious right to if they want.
https://news.ycombinator.com/item?id=32265000 (Yes, social media is undermining democracy)
https://news.ycombinator.com/item?id=25021503 (Ask HN: Instead of picking a party, why can’t we vote on issues)
https://news.ycombinator.com/item?id=21325295 (Ranked-choice voting is on the ballot in New York City)
https://news.ycombinator.com/item?id=12950566 (Maine Ranked Choice Voting Initiative Approved)
https://news.ycombinator.com/item?id=27597662 (What is ranked-choice voting and why is New York using it?)
https://news.ycombinator.com/item?id=24943441 (Maine Becomes First State to Use Ranked-Choice Voting in a Presidential Election)
https://news.ycombinator.com/item?id=31502540 (Ranked-Choice Voting Is More Democratic, Not Less)
https://news.ycombinator.com/item?id=9477747 (Quadratic Voting)
https://news.ycombinator.com/item?id=18391757 (Instant runoff voting is the system we need)
https://news.ycombinator.com/item?id=24561551 (Ranked-choice voting will be used in Maine’s presidential election, court rules)
https://news.ycombinator.com/item?id=2484677 (A “perfect” voting system)
Power should be distributed closer to constituents.
"The powers not delegated to the United States by the Constitution, nor prohibited by it to the States, are reserved to the States respectively, or to the people."
Dead simple. Put a check mark beside the names you support.
That’s it.
Implicit in the voting process is deciding what you think "support" means. For some voters, that threshold will be "I like this candidate, just not as much as my top preference" while other voters will set the bar at "I could live with this candidate winning, as they're not quite as bad as the candidate I'm most afraid of".
Not only does this mean that different voters are effectively answering different questions, but it means voters may feel overwhelmed or confused (as in The Paradox of Choice, and exponentially more so than picking just one top preference).
This also leads to tactical voting, with some politicians or activists telling their supporters to "approve" (or remember not to approve) of some third candidates who they think might gain not enough (or too many) just-in-case votes.
Perhaps these concerns aren't a reason not to try implementing it for elections, but I'd feel more confident about it as a voting method once I've seen how it stands up to large scale malicious campaigning, with groups trying to undermine confidence in the process or exploit voters' unfamiliarity with it.
That is not necessarily desirable though. We already have that a lot of the time (not always), and it seems to systemically produce a race to the bottom.
Having it under FPTP instead of ranked choice is fundamentally different, so we really don't have it at all as far as the FPTP vs RCV debate goes.
What we need are term limits.
Term limits just change the balance away from elected politicians (who can build up a reputation for integrity and competence) and towards the unelected party machinery, who inexperienced politicians are more dependent on.
(Of course, for life-time appointments like supreme court justices, term limits would be an improvement, you're right).
Not only that, but they remain in power with fantasy mandates, because people who hated them were forced to rank them. In Australia, it's criminal not to rank candidates you hate, and criminal to encourage people not to include particular candidates in their rankings, aka a Langer Vote: https://en.wikipedia.org/wiki/Albert_Langer
Australia: Political activist becomes first prisoner of conscience for over 20 years (Albert Langer)
Anything that isn’t simple to understand will fail to get traction. Or worse, will lead to more people questioning the validity of election results.
The American intelligentsia needs to better understand how and why normies feel about things, instead of throwing big words like “ranked choice voting” at them.
Do they, though? Does "largest" mean "tallest" or "heaviest"? Is a tiger bigger than a lion? Do you mean the male of the species or the female?
Sorry if that's strawmanning your argument, but I think that the parent post had a valid point. Ranked choice voting requires that either the voter has a fully formed opinion on all the candidates running, or that they leave some candidates unranked (which makes some people's votes less expressive/powerful than other people's).
Maybe in every election you've taken part in, you've studied the manifestos and voting records of all the candidates, in which case well done and thank you for your civic diligence, but RCV would (hopefully) make it common for a dozen different parties to compete, most of which a given voter would have varying degrees of support for.
Do you really think that the average US voter could memorize an ordering for a list of 12 names?
Perhaps 90% of voters could, and perhaps the remaining 10% of voters don't "deserve" to have their votes fully count, or maybe RCV is such an improvement that this level of disenfranchisement is worth it, but that's not self-evidently true and I don't think the parent post's criticism can be so easily dismissed.
Illegal in Australia. You're required to have an ordered opinion on the relative fitness of every candidate under pain of law, or at least to pretend to have one.
I think they are doing it wrong. RCV is usually "rank the candidates that you know/care about."
In any case, as you say "at least to pretend to have (an opinion)", Australian voters probably just rank the ones they don't know/care about in the order that they appear on the ballot. And since those are probably little-known candidates, it doesn't matter, they aren't getting high votes anyway.
In the children's quiz, it's about objective facts and they're not trying to do "gotchas". The voting will be about opinions.
> Do you really think that the average US voter could memorize an ordering for a list of 12 names?
I expect the voter to show with 1-3 top names, rank them first, then rank everyone else by their preferred party, then the least bad person of the opposition party. I also expect things like pirate/socialist/green/libertarian parties to win elections they otherwise would have been strategic throwaway spoilers in.
> maybe RCV is such an improvement that this level of disenfranchisement is worth it
Another user mentioned something like giving each candidate a 1-5 star rating, I thought that idea was worth exploring.
In ranked choice voting, can I say that one candidate is worthwhile, and all the others would be catastrophic? Can I say that the top five are passable, but the other two are Nazis?