This is actually a really balanced article, making some very nuanced points about why some activists might oppose a system that produces more moderate candidates, and how the existing system sort of relies on low turnout to give extra power to the most motivated core voters of the two main political parties.
It also discusses the problems with Ranked Choice Voting specifically, with examples of its possible failure modes, and considerations of it being harder for less informed or less educated voters to understand, including the issue that not filling in a complete set of preferences means your vote might get discarded before the final round, whereas someone else's wouldn't.
The "issue" of ballot exhaustion is super overblown and preys on people's misunderstandings of instant runoff voting. If you vote for, say, the green party candidate in major American election, your vote is effectively immediately "exhausted" since they will not have even a tenth of the votes of a major party candidate.
FPTP encourages people to metagame who they think is "viable" which is really difficult in unpolled local elections and contrary to people voting who they would actually like to win.
I absolutely agree that FPTP has all the same problems, only worse because they are not made explicitly and transparently part of the voting mechanics.
I still think that ballot exhaustion is a real issue though (maybe an overblown one, but not super overblown). If you vote for the green party candidate, as per your example, but it's an RCV election, and you don't specify your preference for Democrats vs Republicans (because you're only familiar with FPTP, or you feel this would be "disloyal" to the Green candidate) then your vote would be less effective than another Green supporter who did express their full set of preferences.
Maybe it's a price worth paying, that society says "If you don't maximally fill out the ballot, then you are potentially punished by having less political power than some of your fellow citizens", but I think we need to admit that this is a trade-off we're making when we support RCV. (Again, though, it's not a trade-off relative to FPTP, which absolutely gives some voters less power, because votes for third parties are immediately exhausted, I agree).
You're not punished by having less political power, if you want the greens or noone, you get exactly what you want, and if you have a ranked preference, that is available to you. It just has a scary name that makes people think something undemocratic is going on, but nothing is. It's depressing it's explained like you have less political power than your fellow citizens.
If it's not a trade off compared to FPTP, are you comparing to a parliamentary system? How would that work for individual or nonpartisan posts, like say a town mayor or chief dogcatcher?
> if you want the greens or noone, you get exactly what you want
Except you don't get no-one, you get one of the two big parties. (Getting no-one would be like "none of the above" winning, which is a supported outcome in some election systems).
The best I can do to steelman your argument is to imagine you meant "the greens or anyone", because then, yes, the winner was "anyone", but that's not what a voter necessarily thought they were doing by being "loyal" to the green candidate.
Instant-runoff voting (used in practice for counting RCV ballots in the US) actually satisfies the "later-no-harm" criterion[0] of voting systems (unlike Approval Voting), so if a voter knows this fact, they may feel comfortable adding the extra signal of additional preferences to their vote, but otherwise they might be reluctant to, making their vote less effective due to lack of education or misplaced loyalty.
You're right that this is a minor problem compared to the extensive flaws in FPTP, but it is a new and avoidable flaw, and it will affect people who might have previously benefited from (the unfairness of) FPTP. I'm not "comparing to a parliamentary system", though, since both presidential and parliamentary systems are compatible with both RCV/IRV and FPTP, and I think everything we've mentioned is equally applicable to individual and non-partisan posts.
> Instant-runoff voting (used in practice for counting RCV ballots in the US) actually satisfies the "later-no-harm" criterion[0] of voting systems (unlike Approval Voting)
This is a flaw for instant runoff voting at a benefit for approval voting.
In a nutshell, later no harm means voting for your second can't hurt your first. But it can also mean that your vote for your second can't hurt your third either. Thus if it's more important to target your third than your first, you want to hoist your second favorite into the first position by burying your true favorite. That can never be strategically good with approval voting.
You're saying that "later no harm means ... your vote for your second can't hurt your third", which seems strange, given that (by definition) it means your vote for your third can't hurt your second.
If you're right, then your second and third preferences can't harm the chances of each other, and it seems like this result must generalize such that your Nth can't hurt your Mth, for all N and M. That would mean that every possible vote has no effect on the outcome of the election, which seems like a contradiction.
Obviously I'm missing something, but I'm happy to put that down to the complexity and unintuitive outcomes of RCV/IRV, since Approval Voting is a much simpler system to reason about (and implement).
I still disagree that Approval's lack of later-no-harm is a feature, though. In practice, what it means is that voters have to think strategically about how well they think the various candidates will do, in order to decide whether they have to "risk" giving an approval to a candidate who they genuinely don't like, but who might be the only chance of preventing a detested candidate from winning.
i didn't say your 2nd can't hurt your 3rd, but it can't *IF* your 2nd is eliminated before your 1st. which often happens. for instance, a simplified version of what just happened in alaska:
begich is eliminated first, even though he's preferred to each of his rivals by a sizable majority.
but more notably here, palin is a spoiler. her supporters get peltola—their least favorite. BUT if they insincerely/strategically rank begich in 1st instead of palin, then begich wins. i.e. they get their 2nd choice instead of their 3rd.
so they have to LIE ABOUT THEIR FIRST CHOICE in order to prevent the spoiler effect. whereas this is mathematically proven never to happen with approval voting.
> I still disagree that Approval's lack of later-no-harm is a feature, though. In practice, what it means is that voters have to think strategically about how well they think the various candidates will do, in order to decide whether they have to "risk" giving an approval to a candidate who they genuinely don't like, but who might be the only chance of preventing a detested candidate from winning.
you can't argue about strategic voting because:
1. as i just showed you, IRV/RCV has an even more severe vulnerability to this—you have to strategize about whether to support your FAVORITE, not whether to support your 2nd favorite.
2. approval voting gets better results with strategic voters than IRV does with HONEST voters, so even in the worst case scenario, approval voting is still better. see these VSE calculations by harvard stats PhD jameson quinn.
Thank you for clarifying what you meant. Your comment is much easier to read than the article you linked to.
> if they insincerely/strategically rank begich in 1st instead of palin, then begich wins. i.e. they get their 2nd choice instead of their 3rd.
So it sounds like the real complaint here is that IRV fails the "favorite betrayal criterion"[0] which requires that "voters should have no incentive to vote someone else over their favorite".
You're right, that's a genuine flaw of IRV, and one that Approval Voting doesn't suffer from. Unfortunately, the strict definition is slightly unhelpful here, because in Approval Voting the only way to rate a candidate "over" your first preference is to not approve your first preference, which of course will harm them. So it's almost impossible for Approval to fail this criterion just by the lack of options it gives to voters.
Instead we should be considering if voters have an incentive to vote for another candidate over, or as equal to, their first preference. This criterion doesn't have a name, but it's a criterion that Approval doesn't satisfy, for the same reason as your Alaska example. Supporters of Palin have an incentive to approve Begich, i.e. give him equal support on their ballot paper to their actual first preference. You might say that this is fine, because it allows them to get their 2nd preference as the winner instead of their 3rd, but if other people had ended up voting different on the day, then giving this extra support could have denied the Palin supporter the Palin win they were hoping for.
> you can't argue about strategic voting
You're probably right that IRV has more scope for strategy than Approval, and I might even grant that Approval gets better results in practice, regardless of the amount of strategic voters, but I still think that IRV has a better UX and requires less cognitive load (as long as voters don't need to number every candidate, although that makes IRV even less fair).
My point is, as a voter going to fill out your ballot, you know what it means to like one candidate more than another, and you can ignore the candidates you don't know about, and you can probably pick an order randomly if there are two candidates who you like equally well (and randomized ballot ordering helps here, although it potentially hinders illiterate people who need to see the ballot paper long in advance to memorize how to fill it in).
By contrast, being asked which candidates you "approve" of basically can't be answered without having to think strategically, or at least it will mean different things to different people. For some people, "approve" means "I'd still be happy if this person wins", but for others it means "I'd grudgingly accept this person winning as at least they're not the worst candidate on the list". The decision for how to interpret the question will be subject not just to personal psychological biases, but candidate campaigning, and misinformation, and will lead some voters into thinking that they need to pick this "threshold for approval" based on a deep understanding of the polling data (and therefore without that, it might be better not to vote at all).
In conclusion, I don't think this is a reason not to support Approval Voting, especially where the alternative is IRV, but I do think that as electorates get a chance to use it in practice (and campaigns adjust to the new realities it brings), people might start to be annoyed at its flaws. Fortunately those flaws are smaller than those of IRV, and minuscule in comparison to FPTP, but I still think that voting reform advocates can do even better, in terms of UX and simplicity.
is it a real issue if its failure mode in the worst case scenario is the exact same standard case as FPTP? They don't have less political power, they are either choosing or neglecting to use the power they were provided.
I'm failing to see how the person in your example is worse off in any way under RCV than they were in FPTP, and if that's not your implication than what's the point in bringing it up? There's superior voting systems like STAR, but those aren't on the table to be adopted by the country, unlike RCV which already has some support and has been implemented in a few states
I wish these movements for "Ranked Choice Voting" would be recast as RCV/Condorcet (same input, differing tallying), rather than revolving around runoff voting. Runoff voting is essentially still bound to the idea of the two party system, with the added ability for an informational third-party vote in addition to your "real" duopoly choice. Whereas a tallying process that abides by the Condorcet criterion makes sure that whoever gets elected really was actually favored by most voters. The difference will start to matter if third parties ever gain ground.
I agree that a Condorcet counting process would give better outcomes than the IRV one that is usually proposed, you're right, but I think that both are unintuitive, in the sense that if you handed someone a set of ranked ballots and asked them to calculate the winner, they would not be able to come up with a system on their own. In fact, even if you told them what the process should be, they probably wouldn't be able to remember it and implement it correctly.
This lack of transparency/legibility between the voting process and the results it produces is just the sort of thing that undermines people's confidence in elections, and that makes it an easy target for those who oppose reform, or who just want to undermine faith in an outcome that didn't go their way. We only have to look at the 2011 UK referendum on introducing the "Alternative Vote" system (their name for IRV) where the Prime Minister said in the national press that "It’s a puzzling mess of preferences, probabilities and permutations."[0]
I support any reform that is offered to replace FPTP, but advocates need to be really careful that they don't tie themselves to a proposal that is open to such attacks, and then poisons the well for years to come, preventing other more understandable or more representative reforms from being offered (as happened in the UK).
To your point, at least with Condorcet, you can say that the winner was preferred to every other candidate by a majority of people. That isn't a specification of the counting process, but it seems like a pretty straightforward definition of a winner that most everyone can understand.
Runoff voting seems to completely fall to your criticism, in that the outcome can be non-intuitive and really does depend on the counting process. Heck I understand the algorithm and can still foresee examples where it would feel completely unfair.
The real criticism of Condorcet is the possibility of ties. But really that seems like more of a feature, compared to a greedy algorithm that produces non-intuitive outputs. If there is a Condorcet tie, then that should be considered exceptional and a winner should be chosen some other way. Or heck, maybe the winners should be the whole group of 3+ tied candidates, who'd then need a quorum amongst themselves for every action in office.
> They don't have less political power, they are either choosing or neglecting to use the power they were provided.
That sounds like victim blaming. If more people "neglect" to use their power, or if the people who "neglect" to use their power are those who disproportionately prefer one party over another, then you're changing the outcome of elections based on the new voting system's inability to capture the true intent of voters (rather than improving its accuracy).
> if that's not your implication than what's the point in bringing it up?
I think it is still on-topic for this article/discussion to suggest that other reforms would be even better, but I also think that some people under RCV will be worse off than other people, and they will be different people to those who were worse off under FPTP. Like I say, that might be a price worth paying if there are fewer worse-off people under RCV, but we need to be honest with ourselves about that (and aim for voting systems where the fewest possible people have less voting power than their fellow citizens).
> That sounds like victim blaming. If more people "neglect" to use their power, or if the people who "neglect" to use their power are those who disproportionately prefer one party over another, then you're changing the outcome of elections based on the new voting system's inability to capture the true intent of voters (rather than improving its accuracy).
Is there any point where you would concede it was up to the individuals agency? I’m not trying to be combative, but whether or not you think this point exists matters a lot for continuing discourse
If the voting process was made as simple as possible (taking into account the best available evidence from user testing, psychology, linguistics, and so on) and someone decided to deliberately disregard the instructions given, and avoided all efforts to be educated on the topic, then at that point I would agree that the individual's agency was at fault.
To be clear, I don't think that such an individual would "deserve" disenfranchisement as a "punishment" for their actions, since voting is a fundamental right that needs proper due process to remove from someone; it would just be an unfortunate practical consequence of their poor decisions.
Let me reverse the question though: Is there any point where you would concede that the voting system was unacceptably flawed if it caused one or more people to mistakenly fill out a ballot that didn't reflect their true intention (or made them choose not to vote at all)?
> I also think that some people under RCV will be worse off than other people, and they will be different people to those who were worse off under FPTP.
I think I see what you are saying. Essentially the "effort ceiling" for FPTP is very low - everybody gets a single choice. But RCV raises the ceiling, so that high-effort voters have more power than low-effort voters (people who still only put a single choice). This gives advantage to people who have the time and resources to research candidates more thoroughly and make more detailed votes
Two parties is a feature often defined by State election law. Some States like Illinois specifically mention Rs and Ds. Other States often use a method of favoring (donation limits) major parties which is determined by off-year results in a race like the Governor.
Partisanship matters because of benefits in State election laws. State legislators who are Ds and Rs know this and don't talk about it because it would reduce their club powers.
I would clean out the State election laws before applying band-aids and lotions.
The US two party oligarchy is particularly silly given the poor unrepresentative outcomes and the vocal opposition to political parties overall by many of the US founding fathers.
> I would clean out the State election laws before applying band-aids and lotions.
I don't see why the two approaches can't be run in parallel. I agree with you, though, that laws which privilege specific parties (or even parties generally over independent candidates) have no place in a democracy.
Have there not been any cases brought which challenge the constitutionality of such laws?
I know that Lawrence Lessig tried to challenge Winner Take All (plurality) voting in the US under the Equal Protection Clause of the 14th Amendment[0], which seemed like a bit of a long shot, whereas I don't think that the establishment would feel particularly threatened by having to rewrite a few state laws to be more neutral. Presumably there are plenty of states that don't have such laws, and they are equally locked into the two-party system.
Sometimes difficult to show harm and standing. Appeals are expensive.
The States I looked at have major/minor party and other hindrances such as mandatory reports to secretary of state by party officers from each county. Maybe I would look in N.H. election laws for a place without this.
> and considerations of it being harder for less informed or less educated voters to understand,
Works fine in Australia and I'll add, in the true spirit of Aussie Competiveness, that our dumb F*'s are dumber than US dumb F*'s.
Making a list of who you like in numbered order ain't that hard, and ditching the ballots of those that cannot number 1 through 6 is no great loss.
The USofA has one of the truly insanely over complicated and least effective voting systems on the planet .. the billions spent on presidential elections is just ridiculous and symptomatic of some kind of crazy "little king" syndrome when the real focus should be on getting a balanced representative team together .. let them sort out a token figurehead that can be easily replaced if corrupt | sick | etc.
> ditching the ballots of those that cannot number 1 through 6 is no great loss.
It's worth looking into the statistics around that. For example, one study[0] found that:
"Between 2004 – 2016, 32% of electorates reported more informal votes than votes in the margin between the winner and runner-up."
Of course not, not all "informal votes" are due to people trying but failing to correctly fill out their ballot paper, but the study also found that "more candidates cause more informal votes", so the complexity of the ballots was clearly a factor.
The fact that as many as a third of seats would have a different winner if every vote was counted, though, has to undermine confidence in the democratic process. And this is in a polity where voting is compulsory, so in the US it would be more likely that these informal voters just stayed at home, and the problem would be even more brushed under the carpet.
An interesting study, cheers for that, a few points of note that sprang out at me:
- In Australia, where voting is compulsory, around 5% of votes are informal, not counting toward the outcome.
So, compulsory voting (> 95% of those that can vote do vote) plus low (5%) informal voting ==> way more people vote effectively in Australia than in the USofA (by a long shot).
- The abstract quote differs from the main body quote of the same:
> Between 2004–2016, 32% of electorates received more informal votes than those in the margin between the winner and runner-up.
Or, in other words: 32% of electorates were closely contested and could have been swayed by the ineffective votes of those that didn't pay attention to detail (is this what we want?).
What isn't expanded on (in my quick skim, I could be wrong) - is that these electorates were close AFTER ranking votes were counted and less popular candidates discarded with second and third preferences thrown in.
Your "conclusion"
> The fact that as many as a third of seats would have a different winner if every vote was counted, though, has to undermine confidence in the democratic process.
is hasty - it assumes that the informal votes, had they been proper votes (rather than drawings of a dick, etc) would have made a decisive and coordinated swing bloc large enough to tip the balance .. when, in truth, they could have been evenly split across the field and had no decisive weight.
(And there are other issues, but that'll suffice for now).
Thanks for digging into that study and providing the extra context.
You make a great point that Australian elections, even with the informal votes, have a greater quantity of voter expressions, and also a greater quality of expressions, since the ranking provides more fine-grained data about the voter's wishes (for those voters who can effectively wield that vote). That is the beginning of quite a compelling argument that the Australian electoral system probably, in practice, produces better outcomes than FPTP in the US, even with the additional losses it incurs due to its complexity.
I also admit that we don't really have data on how many of those 32% of electorates actually would be affected in practice, for the reasons you mention (which is why I went with the slightly "safe" language of "as many as a third of seats"), since this is an upper bound, but it is a worryingly high upper bound for something we want the public to have confidence in.
What I probably disagree with most in your comment is the dismissive or derisive tone you take towards "those that didn't pay attention to detail". On a forum like this, it's probably a relatively popular idea that those who are the most intelligent and well-informed and have the best attention to detail, should have the most say in running the country, but it's actually a dangerously corrosive idea. Instead, we should be using our intelligence to design systems that are resilient to user error, and that have the maximal probability of picking the correct winner, while putting the smallest cognitive burden on the voter, who should be entitled to equal protection and equal representation under the law (even if those ideals are far from being met in multiple aspects of civic life).
(tl;dr: If you're in a state with closed and/or partisan primaries, that's probably the easiest and highest-impact change - and doesn't require changing voting methods. https://ballotpedia.org/Primary_election_types_by_state has more.)
The article discusses two independent changes together:
1. Changing to open, non-partisan primaries (where every voter can vote for any candidate in the primary), and
2. Changing the method by which the winner is chosen.
In terms of making winners better represent the entire electorate, moving to a non-partisan primaries might be the single highest-impact improvement that a jurisdiction can make. It almost certainly is when one considers the chances of the change actually happening (ie, viability).
Separately, jurisdictions can elect to change voting methods in elections with more than two candidates. For example, I’m co-leading a campaign to change Seattle's open primaries to use Approval Voting[1], which is on the Seattle ballot as Proposition 1A[2] now.
My point is that each of these changes has supporters and detractors, pros and cons. Acting like they both need to be changed at once makes things harder.
I remember in the Atlanta area when Cynthia McKinney was running for reelection, and a large number of conservatives crossed party lines to vote for her opponent, Denise Majette. It upset a lot of Democrats, but the idea of cross primary voting seems to be one that both parties used to undermine the other.
That would seem a strike against an open system, especially if you have a party with a larger presence in a district that can negatively disrupt the other party.
Of course the idea is to have no parties, but like-minded people coalesce, and it’s nearly impossible to get rid of them. You could try nonpartisan, but with nearly 200 years of a party system, it would be nearly impossible to ever get rid of it. Without infringing on free speech of course
> it would be nearly impossible to ever get rid of it. Without infringing on free speech of course.
Most State election laws have major/minor party designations which hinder the minor parties. Removing those State law designations would not affect free speech.
Maybe you can search your State election law and let us know what you find about things such as donation limits to minor parties.
> Instead, Nevada would have a nonpartisan primary, from which the top five candidates of any party would emerge to the general election.
How useful is this? My understanding in that in a fair number of elections there is no challenger to the incumbent party, or the challenge is token at best.
That is the two parties have implied territories, sans a handful of areas that are up for grabs "battlegrounds".
With traditional voting or with ranked voting, I'd like to see "None of the above" be provided as a candidate. The People should be provided a means to express that feeling if that's the case.
Let be honest, low voter turnout is often a function of the products being offered. "Not of the above" provides important feedback that's currently lost.
"Null" is not tracked. Also, it cannot be assumed to be intentional. It's the lack of a vote. NotA is a decision and actively communicates dissatisfaction with the choices offered.
You can't show up, leave everyhing blank, and presume that void will be heard.
Spoilt / informal ballots are counted in elections in many jurisdictions, with the number of such votes reported in the final results table. This means that there is at least an upper-bound known for the number of people who scribbled "none of the above" across their ballot.
If the NotA campaign were more organized, and supporters of it were more motivated, then they could probably get the number of spoilt / informal ballots well above the margin of victory in multiple districts, which would be an effective way of bringing the legitimacy of the candidate's win into question, and pressuring politicians to make NotA an actual option on ballots.
Once that goal was achieved, though, and NotA still got less than 5% of the vote, I think it would be very hard to convince the public that the winner was somehow hiding their true unpopularity by "forcing" voters to vote for someone they didn't like (and that the winner was disingenuously claiming that low turnout meant that the abstaining voters were confident they would be happy with the outcome).
However, because of the winner take all system, small parties can never gain power. You will almost always end up with two parties in a plurality vote since that is the only way to reliably win.
In other words the party with the broadest appeal will always win. If two right wing candidates both receive 30% of the vote, and the single left wing candidate receives 40% of the vote then the left wins power. So in the next cycle, you have an obvious choice: split the right wing vote and lose, or coalesce into a single party and get 60% of the vote.
Look up Maurice Duverger. He did a lot of theoretical work on this.
If I think abortion is fine but spiraling deficits are not am I left wing or right wing?
I suggest an alternative stratagem: partner up with a righty/lefty on orthogonal issues, split both votes, and then collaborate honorably once one of you is in power. As if you are in cahoots but openly. Like a coalition of some sort.
It was a simplified example to explain the game theory behind it.
Abortion/deficits is a result of that. You find (or create) wedge issues that act as an issue that voters won’t budge on, thereby deciding their vote.
If 50% or more of the population feels strongly enough about a given issue that they will always vote based on it then you position yourself as the party that supports that issue and you have a guaranteed win.
I don’t like the system, just explaining that those are the incentives that the rules create.
You know which US party is on the side of "abortion is fine."
As to "spiraling deficits", take a look at the US deficit per year[0]. Where the Bush Administration caused the massive doozy of the 2008 recession where deficits spiked. And then the Obama Administration, upon inheriting that mess, reduced the deficit every single year until 2016. Then, the deficit increased every year under the Trump administration.
You would be Democratic party "left wing" by both accounts.
To add to your point, the U.S. government suffered budget deficits every year from 1970 through 1997, then when Bill Clinton was president in 1998, the government finally recorded a surplus. There were also budget surpluses in 1999, 2000 and in 2001. 2001 was the last year the Clinton administration proposed the budget. Republican George W. Bush succeeded Clinton in 2001, and the US went back to having a budget deficit in 2002.
That's a bit of a coincidence. Bill Clinton presided over the original dotcom boom, which produced a ton of federal tax revenue without any actual work on the government's part. Bush Jr's deficits are due in large part to an economic crash that actually begun under Clinton, as the dotcom boom failed.
Which doesn't really undercut the point: Republicans aren't actually very good at deficit reduction. Deficit reduction requires either increased revenue or decreased spending. They never substantially decrease spending (they always focus on a few small programs they're ideologically opposed to) and prominently run to decrease revenue. Their supposition that lower taxes will somehow increase revenue has never once worked in practice and was always ridiculous in theory.
Democrats don't generally cut the deficit either, except for Bill Clinton's fortunate streak. But at least they don't pretend otherwise, and they don't present it as a tradeoff of "Sure, we're barbaric on human rights issues, but at least we'll lower your taxes while magically decreasing the deficit."
The solution to this is federalism - deconcentrated power at the state level - having one approach to lord over all - especially for a country as large and diverse as the US is asinine. The sooner we go back to federalism and neuter the federal government, the better.
It's a poor law that doesn't have exceptions; in this case the exceptions are many other countries in the G20 .. eg: Australia, three major parties, numerous minor parties and at any time in the past several decades there has been sitting houses with members from more than two parties.
I guess that puts a ding in "mathematically impossible" (given the stability of the Australian political system for 100+ years).
I can't help but feel that the author's point is, "Wouldn't it be great if all states used RCV for primaries?", which while probably an improvement, I am not convinced will change much.
RCV + first past the post(FPTP) still produces two-party systems. They simply make coordination more difficult. In effect, they chose winners based on intra-party voter coordination ability ie: the ability for a party to effectively communicate a preferred ranking. That may or may not be more aligned with your desires than whatever metric we're functionally choosing winners by today.
Still, RCV is an incremental, and I'd argue necessary, step toward multi-winner elections. To that effect, it's most likely an improvement.
That might work. Another path is thru the Senate. It would take only a relative handful of democrat people (say 100,000) to move their voting state from California to North Dakota, South Dakota and Wyoming and vote in Democratic Senators.
(This is from, possibly dodgy, memory, The Economist had an article on this about six months ago that had a detailed plan with all the numbers.)
Couldn't we just solve this by eliminating draconian ballot access provisions? If all 330 million Americans want to run for president I don't think it's a serious logistical issue anymore. Just assign every candidate a registration number and have them put it on all their campaigns ads
An argument in favor of partisan primaries is that it allows parties to pick the candidates that best represent the party platform. Supposedly.
The discussion that is more cogent in my opinion is ballot security.
This is not only the chain of custody of ballots throughout the life cycle, but also the security of the voter marking and casting the ballot. Mail-in ballots offer vast opportunity for voter coercion. Which is why a polling place is really the proper answer for almost all voters.
This is both a huge issue for election credibility, and something that the less serious voices are entirely too glib about.
54 comments
[ 3.3 ms ] story [ 115 ms ] threadIt also discusses the problems with Ranked Choice Voting specifically, with examples of its possible failure modes, and considerations of it being harder for less informed or less educated voters to understand, including the issue that not filling in a complete set of preferences means your vote might get discarded before the final round, whereas someone else's wouldn't.
FPTP encourages people to metagame who they think is "viable" which is really difficult in unpolled local elections and contrary to people voting who they would actually like to win.
I still think that ballot exhaustion is a real issue though (maybe an overblown one, but not super overblown). If you vote for the green party candidate, as per your example, but it's an RCV election, and you don't specify your preference for Democrats vs Republicans (because you're only familiar with FPTP, or you feel this would be "disloyal" to the Green candidate) then your vote would be less effective than another Green supporter who did express their full set of preferences.
Maybe it's a price worth paying, that society says "If you don't maximally fill out the ballot, then you are potentially punished by having less political power than some of your fellow citizens", but I think we need to admit that this is a trade-off we're making when we support RCV. (Again, though, it's not a trade-off relative to FPTP, which absolutely gives some voters less power, because votes for third parties are immediately exhausted, I agree).
If it's not a trade off compared to FPTP, are you comparing to a parliamentary system? How would that work for individual or nonpartisan posts, like say a town mayor or chief dogcatcher?
BTW, it's no one or no-one.
Except you don't get no-one, you get one of the two big parties. (Getting no-one would be like "none of the above" winning, which is a supported outcome in some election systems).
The best I can do to steelman your argument is to imagine you meant "the greens or anyone", because then, yes, the winner was "anyone", but that's not what a voter necessarily thought they were doing by being "loyal" to the green candidate.
Instant-runoff voting (used in practice for counting RCV ballots in the US) actually satisfies the "later-no-harm" criterion[0] of voting systems (unlike Approval Voting), so if a voter knows this fact, they may feel comfortable adding the extra signal of additional preferences to their vote, but otherwise they might be reluctant to, making their vote less effective due to lack of education or misplaced loyalty.
You're right that this is a minor problem compared to the extensive flaws in FPTP, but it is a new and avoidable flaw, and it will affect people who might have previously benefited from (the unfairness of) FPTP. I'm not "comparing to a parliamentary system", though, since both presidential and parliamentary systems are compatible with both RCV/IRV and FPTP, and I think everything we've mentioned is equally applicable to individual and non-partisan posts.
[0] https://en.wikipedia.org/wiki/Later-no-harm_criterion
This is a flaw for instant runoff voting at a benefit for approval voting.
In a nutshell, later no harm means voting for your second can't hurt your first. But it can also mean that your vote for your second can't hurt your third either. Thus if it's more important to target your third than your first, you want to hoist your second favorite into the first position by burying your true favorite. That can never be strategically good with approval voting.
https://medium.com/@ClayShentrup/later-no-harm-72c44e145510
If you're right, then your second and third preferences can't harm the chances of each other, and it seems like this result must generalize such that your Nth can't hurt your Mth, for all N and M. That would mean that every possible vote has no effect on the outcome of the election, which seems like a contradiction.
Obviously I'm missing something, but I'm happy to put that down to the complexity and unintuitive outcomes of RCV/IRV, since Approval Voting is a much simpler system to reason about (and implement).
I still disagree that Approval's lack of later-no-harm is a feature, though. In practice, what it means is that voters have to think strategically about how well they think the various candidates will do, in order to decide whether they have to "risk" giving an approval to a candidate who they genuinely don't like, but who might be the only chance of preventing a detested candidate from winning.
35% peltola begich
33% palin begich
18% begich peltola
14% begich palin
-- https://electionscience.org/commentary-analysis/rcv-fools-pa...
begich is eliminated first, even though he's preferred to each of his rivals by a sizable majority.
but more notably here, palin is a spoiler. her supporters get peltola—their least favorite. BUT if they insincerely/strategically rank begich in 1st instead of palin, then begich wins. i.e. they get their 2nd choice instead of their 3rd.
so they have to LIE ABOUT THEIR FIRST CHOICE in order to prevent the spoiler effect. whereas this is mathematically proven never to happen with approval voting.
> I still disagree that Approval's lack of later-no-harm is a feature, though. In practice, what it means is that voters have to think strategically about how well they think the various candidates will do, in order to decide whether they have to "risk" giving an approval to a candidate who they genuinely don't like, but who might be the only chance of preventing a detested candidate from winning.
you can't argue about strategic voting because:
1. as i just showed you, IRV/RCV has an even more severe vulnerability to this—you have to strategize about whether to support your FAVORITE, not whether to support your 2nd favorite.
2. approval voting gets better results with strategic voters than IRV does with HONEST voters, so even in the worst case scenario, approval voting is still better. see these VSE calculations by harvard stats PhD jameson quinn.
https://electionscience.github.io/vse-sim/VSEbasic/
my link already discussed all of this quite clearly.
> if they insincerely/strategically rank begich in 1st instead of palin, then begich wins. i.e. they get their 2nd choice instead of their 3rd.
So it sounds like the real complaint here is that IRV fails the "favorite betrayal criterion"[0] which requires that "voters should have no incentive to vote someone else over their favorite".
You're right, that's a genuine flaw of IRV, and one that Approval Voting doesn't suffer from. Unfortunately, the strict definition is slightly unhelpful here, because in Approval Voting the only way to rate a candidate "over" your first preference is to not approve your first preference, which of course will harm them. So it's almost impossible for Approval to fail this criterion just by the lack of options it gives to voters.
Instead we should be considering if voters have an incentive to vote for another candidate over, or as equal to, their first preference. This criterion doesn't have a name, but it's a criterion that Approval doesn't satisfy, for the same reason as your Alaska example. Supporters of Palin have an incentive to approve Begich, i.e. give him equal support on their ballot paper to their actual first preference. You might say that this is fine, because it allows them to get their 2nd preference as the winner instead of their 3rd, but if other people had ended up voting different on the day, then giving this extra support could have denied the Palin supporter the Palin win they were hoping for.
> you can't argue about strategic voting
You're probably right that IRV has more scope for strategy than Approval, and I might even grant that Approval gets better results in practice, regardless of the amount of strategic voters, but I still think that IRV has a better UX and requires less cognitive load (as long as voters don't need to number every candidate, although that makes IRV even less fair).
My point is, as a voter going to fill out your ballot, you know what it means to like one candidate more than another, and you can ignore the candidates you don't know about, and you can probably pick an order randomly if there are two candidates who you like equally well (and randomized ballot ordering helps here, although it potentially hinders illiterate people who need to see the ballot paper long in advance to memorize how to fill it in).
By contrast, being asked which candidates you "approve" of basically can't be answered without having to think strategically, or at least it will mean different things to different people. For some people, "approve" means "I'd still be happy if this person wins", but for others it means "I'd grudgingly accept this person winning as at least they're not the worst candidate on the list". The decision for how to interpret the question will be subject not just to personal psychological biases, but candidate campaigning, and misinformation, and will lead some voters into thinking that they need to pick this "threshold for approval" based on a deep understanding of the polling data (and therefore without that, it might be better not to vote at all).
In conclusion, I don't think this is a reason not to support Approval Voting, especially where the alternative is IRV, but I do think that as electorates get a chance to use it in practice (and campaigns adjust to the new realities it brings), people might start to be annoyed at its flaws. Fortunately those flaws are smaller than those of IRV, and minuscule in comparison to FPTP, but I still think that voting reform advocates can do even better, in terms of UX and simplicity.
[0] https://electowiki.org/wiki/Favorite_betrayal_criterion
I'm failing to see how the person in your example is worse off in any way under RCV than they were in FPTP, and if that's not your implication than what's the point in bringing it up? There's superior voting systems like STAR, but those aren't on the table to be adopted by the country, unlike RCV which already has some support and has been implemented in a few states
This lack of transparency/legibility between the voting process and the results it produces is just the sort of thing that undermines people's confidence in elections, and that makes it an easy target for those who oppose reform, or who just want to undermine faith in an outcome that didn't go their way. We only have to look at the 2011 UK referendum on introducing the "Alternative Vote" system (their name for IRV) where the Prime Minister said in the national press that "It’s a puzzling mess of preferences, probabilities and permutations."[0]
I support any reform that is offered to replace FPTP, but advocates need to be really careful that they don't tie themselves to a proposal that is open to such attacks, and then poisons the well for years to come, preventing other more understandable or more representative reforms from being offered (as happened in the UK).
[0] https://www.dailymail.co.uk/debate/article-1383675/AV-refere...
Runoff voting seems to completely fall to your criticism, in that the outcome can be non-intuitive and really does depend on the counting process. Heck I understand the algorithm and can still foresee examples where it would feel completely unfair.
The real criticism of Condorcet is the possibility of ties. But really that seems like more of a feature, compared to a greedy algorithm that produces non-intuitive outputs. If there is a Condorcet tie, then that should be considered exceptional and a winner should be chosen some other way. Or heck, maybe the winners should be the whole group of 3+ tied candidates, who'd then need a quorum amongst themselves for every action in office.
That sounds like victim blaming. If more people "neglect" to use their power, or if the people who "neglect" to use their power are those who disproportionately prefer one party over another, then you're changing the outcome of elections based on the new voting system's inability to capture the true intent of voters (rather than improving its accuracy).
> if that's not your implication than what's the point in bringing it up?
I think it is still on-topic for this article/discussion to suggest that other reforms would be even better, but I also think that some people under RCV will be worse off than other people, and they will be different people to those who were worse off under FPTP. Like I say, that might be a price worth paying if there are fewer worse-off people under RCV, but we need to be honest with ourselves about that (and aim for voting systems where the fewest possible people have less voting power than their fellow citizens).
Is there any point where you would concede it was up to the individuals agency? I’m not trying to be combative, but whether or not you think this point exists matters a lot for continuing discourse
To be clear, I don't think that such an individual would "deserve" disenfranchisement as a "punishment" for their actions, since voting is a fundamental right that needs proper due process to remove from someone; it would just be an unfortunate practical consequence of their poor decisions.
Let me reverse the question though: Is there any point where you would concede that the voting system was unacceptably flawed if it caused one or more people to mistakenly fill out a ballot that didn't reflect their true intention (or made them choose not to vote at all)?
I think I see what you are saying. Essentially the "effort ceiling" for FPTP is very low - everybody gets a single choice. But RCV raises the ceiling, so that high-effort voters have more power than low-effort voters (people who still only put a single choice). This gives advantage to people who have the time and resources to research candidates more thoroughly and make more detailed votes
Two parties is a feature often defined by State election law. Some States like Illinois specifically mention Rs and Ds. Other States often use a method of favoring (donation limits) major parties which is determined by off-year results in a race like the Governor.
Partisanship matters because of benefits in State election laws. State legislators who are Ds and Rs know this and don't talk about it because it would reduce their club powers. I would clean out the State election laws before applying band-aids and lotions.
I don't see why the two approaches can't be run in parallel. I agree with you, though, that laws which privilege specific parties (or even parties generally over independent candidates) have no place in a democracy.
Have there not been any cases brought which challenge the constitutionality of such laws?
I know that Lawrence Lessig tried to challenge Winner Take All (plurality) voting in the US under the Equal Protection Clause of the 14th Amendment[0], which seemed like a bit of a long shot, whereas I don't think that the establishment would feel particularly threatened by having to rewrite a few state laws to be more neutral. Presumably there are plenty of states that don't have such laws, and they are equally locked into the two-party system.
[0] https://medium.lessig.org/the-equal-protection-challenge-to-...
The States I looked at have major/minor party and other hindrances such as mandatory reports to secretary of state by party officers from each county. Maybe I would look in N.H. election laws for a place without this.
That's true, but it seems like an easier case to bring than the cases that have challenged gerrymandering, for example.
Works fine in Australia and I'll add, in the true spirit of Aussie Competiveness, that our dumb F*'s are dumber than US dumb F*'s.
Making a list of who you like in numbered order ain't that hard, and ditching the ballots of those that cannot number 1 through 6 is no great loss.
The USofA has one of the truly insanely over complicated and least effective voting systems on the planet .. the billions spent on presidential elections is just ridiculous and symptomatic of some kind of crazy "little king" syndrome when the real focus should be on getting a balanced representative team together .. let them sort out a token figurehead that can be easily replaced if corrupt | sick | etc.
It's worth looking into the statistics around that. For example, one study[0] found that:
"Between 2004 – 2016, 32% of electorates reported more informal votes than votes in the margin between the winner and runner-up."
Of course not, not all "informal votes" are due to people trying but failing to correctly fill out their ballot paper, but the study also found that "more candidates cause more informal votes", so the complexity of the ballots was clearly a factor.
The fact that as many as a third of seats would have a different winner if every vote was counted, though, has to undermine confidence in the democratic process. And this is in a polity where voting is compulsory, so in the US it would be more likely that these informal voters just stayed at home, and the problem would be even more brushed under the carpet.
[0] http://eamonmcginn.com.s3-website-ap-southeast-2.amazonaws.c...
- In Australia, where voting is compulsory, around 5% of votes are informal, not counting toward the outcome.
So, compulsory voting (> 95% of those that can vote do vote) plus low (5%) informal voting ==> way more people vote effectively in Australia than in the USofA (by a long shot).
- The abstract quote differs from the main body quote of the same:
> Between 2004–2016, 32% of electorates received more informal votes than those in the margin between the winner and runner-up.
Or, in other words: 32% of electorates were closely contested and could have been swayed by the ineffective votes of those that didn't pay attention to detail (is this what we want?).
What isn't expanded on (in my quick skim, I could be wrong) - is that these electorates were close AFTER ranking votes were counted and less popular candidates discarded with second and third preferences thrown in.
Your "conclusion" > The fact that as many as a third of seats would have a different winner if every vote was counted, though, has to undermine confidence in the democratic process.
is hasty - it assumes that the informal votes, had they been proper votes (rather than drawings of a dick, etc) would have made a decisive and coordinated swing bloc large enough to tip the balance .. when, in truth, they could have been evenly split across the field and had no decisive weight.
(And there are other issues, but that'll suffice for now).
You make a great point that Australian elections, even with the informal votes, have a greater quantity of voter expressions, and also a greater quality of expressions, since the ranking provides more fine-grained data about the voter's wishes (for those voters who can effectively wield that vote). That is the beginning of quite a compelling argument that the Australian electoral system probably, in practice, produces better outcomes than FPTP in the US, even with the additional losses it incurs due to its complexity.
I also admit that we don't really have data on how many of those 32% of electorates actually would be affected in practice, for the reasons you mention (which is why I went with the slightly "safe" language of "as many as a third of seats"), since this is an upper bound, but it is a worryingly high upper bound for something we want the public to have confidence in.
What I probably disagree with most in your comment is the dismissive or derisive tone you take towards "those that didn't pay attention to detail". On a forum like this, it's probably a relatively popular idea that those who are the most intelligent and well-informed and have the best attention to detail, should have the most say in running the country, but it's actually a dangerously corrosive idea. Instead, we should be using our intelligence to design systems that are resilient to user error, and that have the maximal probability of picking the correct winner, while putting the smallest cognitive burden on the voter, who should be entitled to equal protection and equal representation under the law (even if those ideals are far from being met in multiple aspects of civic life).
The article discusses two independent changes together:
1. Changing to open, non-partisan primaries (where every voter can vote for any candidate in the primary), and
2. Changing the method by which the winner is chosen.
A jurisdiction can move to open, non-partisan primaries without a new voting method. As the article mentioned, Washington did so: https://ballotpedia.org/Primary_elections_in_Washington . Other states: https://ballotpedia.org/Primary_election_types_by_state
In terms of making winners better represent the entire electorate, moving to a non-partisan primaries might be the single highest-impact improvement that a jurisdiction can make. It almost certainly is when one considers the chances of the change actually happening (ie, viability).
Separately, jurisdictions can elect to change voting methods in elections with more than two candidates. For example, I’m co-leading a campaign to change Seattle's open primaries to use Approval Voting[1], which is on the Seattle ballot as Proposition 1A[2] now.
My point is that each of these changes has supporters and detractors, pros and cons. Acting like they both need to be changed at once makes things harder.
[1]: https://electionscience.org/approval-voting-faqs/
[2]: https://seattleapproves.org/
That would seem a strike against an open system, especially if you have a party with a larger presence in a district that can negatively disrupt the other party.
Of course the idea is to have no parties, but like-minded people coalesce, and it’s nearly impossible to get rid of them. You could try nonpartisan, but with nearly 200 years of a party system, it would be nearly impossible to ever get rid of it. Without infringing on free speech of course
Most State election laws have major/minor party designations which hinder the minor parties. Removing those State law designations would not affect free speech. Maybe you can search your State election law and let us know what you find about things such as donation limits to minor parties.
How useful is this? My understanding in that in a fair number of elections there is no challenger to the incumbent party, or the challenge is token at best.
That is the two parties have implied territories, sans a handful of areas that are up for grabs "battlegrounds".
With traditional voting or with ranked voting, I'd like to see "None of the above" be provided as a candidate. The People should be provided a means to express that feeling if that's the case.
Let be honest, low voter turnout is often a function of the products being offered. "Not of the above" provides important feedback that's currently lost.
You can't show up, leave everyhing blank, and presume that void will be heard.
If the NotA campaign were more organized, and supporters of it were more motivated, then they could probably get the number of spoilt / informal ballots well above the margin of victory in multiple districts, which would be an effective way of bringing the legitimacy of the candidate's win into question, and pressuring politicians to make NotA an actual option on ballots.
Once that goal was achieved, though, and NotA still got less than 5% of the vote, I think it would be very hard to convince the public that the winner was somehow hiding their true unpopularity by "forcing" voters to vote for someone they didn't like (and that the winner was disingenuously claiming that low turnout meant that the abstaining voters were confident they would be happy with the outcome).
https://www.commoncause.org/colorado/democracy-wire/first-pa...
There_I_fixed_it.jpeg
I do get why the incumbent parties wouldn't want to change the system but surely if day dreaming about saving America this would be an area of focus.
However, because of the winner take all system, small parties can never gain power. You will almost always end up with two parties in a plurality vote since that is the only way to reliably win.
In other words the party with the broadest appeal will always win. If two right wing candidates both receive 30% of the vote, and the single left wing candidate receives 40% of the vote then the left wins power. So in the next cycle, you have an obvious choice: split the right wing vote and lose, or coalesce into a single party and get 60% of the vote.
Look up Maurice Duverger. He did a lot of theoretical work on this.
I suggest an alternative stratagem: partner up with a righty/lefty on orthogonal issues, split both votes, and then collaborate honorably once one of you is in power. As if you are in cahoots but openly. Like a coalition of some sort.
Abortion/deficits is a result of that. You find (or create) wedge issues that act as an issue that voters won’t budge on, thereby deciding their vote.
If 50% or more of the population feels strongly enough about a given issue that they will always vote based on it then you position yourself as the party that supports that issue and you have a guaranteed win.
I don’t like the system, just explaining that those are the incentives that the rules create.
You know which US party is on the side of "abortion is fine."
As to "spiraling deficits", take a look at the US deficit per year[0]. Where the Bush Administration caused the massive doozy of the 2008 recession where deficits spiked. And then the Obama Administration, upon inheriting that mess, reduced the deficit every single year until 2016. Then, the deficit increased every year under the Trump administration.
You would be Democratic party "left wing" by both accounts.
[0]https://www.thebalancemoney.com/us-deficit-by-year-3306306
https://politicalticker.blogs.cnn.com/2010/02/03/cnn-fact-ch...
Which doesn't really undercut the point: Republicans aren't actually very good at deficit reduction. Deficit reduction requires either increased revenue or decreased spending. They never substantially decrease spending (they always focus on a few small programs they're ideologically opposed to) and prominently run to decrease revenue. Their supposition that lower taxes will somehow increase revenue has never once worked in practice and was always ridiculous in theory.
Democrats don't generally cut the deficit either, except for Bill Clinton's fortunate streak. But at least they don't pretend otherwise, and they don't present it as a tradeoff of "Sure, we're barbaric on human rights issues, but at least we'll lower your taxes while magically decreasing the deficit."
I guess that puts a ding in "mathematically impossible" (given the stability of the Australian political system for 100+ years).
RCV + first past the post(FPTP) still produces two-party systems. They simply make coordination more difficult. In effect, they chose winners based on intra-party voter coordination ability ie: the ability for a party to effectively communicate a preferred ranking. That may or may not be more aligned with your desires than whatever metric we're functionally choosing winners by today.
Still, RCV is an incremental, and I'd argue necessary, step toward multi-winner elections. To that effect, it's most likely an improvement.
(This is from, possibly dodgy, memory, The Economist had an article on this about six months ago that had a detailed plan with all the numbers.)
The discussion that is more cogent in my opinion is ballot security.
This is not only the chain of custody of ballots throughout the life cycle, but also the security of the voter marking and casting the ballot. Mail-in ballots offer vast opportunity for voter coercion. Which is why a polling place is really the proper answer for almost all voters.
This is both a huge issue for election credibility, and something that the less serious voices are entirely too glib about.