My vote on voting systems
blog.danielh.cc
blog.danielh.cc
Voting is even more about identifying who didn't win than who did win. In a multi-member district a much smaller slice of the electorate 'loses', making it much easier to build broad consensus around results.
Some of the sharper edges of many ranked choice or scoring voting systems are blunted by multi-member districts. They are not winner-take-all, so small changes in expressed preference that would have changed the winner in a single member district just re-orders the winners. Specific candidates could still find themselves losing when another system would have elected them, but the electorate as a whole still ends up with representation they can tolerate.
But you still need to work around collusions between candidates. Candidates will be pushed to run on a slate of N candidates for N seats, and promise to support each other. Voters who want any of their positions will avoid somebody who would interfere with their own top priority. The result acts the same as a two party system, where all N candidates support the same party.
You can get some benefit by combining it with List Voting, where you go ahead and reify the party system. The party gets to send whoever it wants, proportional to the votes.That makes it harder to run multiple candidates that take all of the spots. (Not impossible: they can break into multiple virtual parties. But that's risky and coordinating votes is difficult.)
Even at that I'm not sure it really helps. Lawmaking is still fundamentally a "single member" operation: a bill either passes or it doesn't. You still get pushed to be in either the majority or minority coalition.
The solution for that is to not make it a simple binary pass/fail, but to have how it passes and fails affect the outcome.
A simple and consistent way of doing this is by making it so that all laws have a built in expiration date, and how big the majority for passing the bill affects how long the law is in effect from.
A 50% + 1 margin of victory should result in the law becoming void the moment the parliament composition changes.
Perhaps a 60% margin might allow it to go for 10 years, a 70% margin for 15 years, and so on.
Constitutional changes should essentially require unanimity ( and a referendum on top of it) since they are the only permanent laws.
Some laws would be more or less "permanent" since every 30 years or so you'd like get a 99%+ majority for a "theft is illegal" type low, but other more controversial would likely be on the ballots every few years.
I do like the idea of built in timers on laws, but instant revocation seems too extreme. A marginal law expiring at the end of the new term seems more conducive to stability. That gives time to debate, discuss, and pass (or not) a modified version so everyone affected has time to participate and plan for revisions.
> Constitutional changes should essentially require unanimity ( and a referendum on top of it) since they are the only permanent laws.
Here in Canada the bar on constitutional amendments has been set so high it's unlikely we will ever be able to change it for any reason. Too much change isn't healthy, but neither is none.
Not American, but I do think their bar for constitutional amendments has also been set too high and I think it's contributing to polarization. Setting it even higher seems counter-productive. I have similar feelings about the filibuster.
More laws passing would let people live with the consequences, good or bad, of all the stuff being proposed. Voters would be able to vote based on reality rather than the boogey-man each side portrays of what the other side might do with power. There should be a high bar to change a constitution, but it needs to be clearable more often.
Candidate-centered proportional systems (e.g., single transferrable vote in moderately sized, 5-7 member, districts) do not have this problem of party-list/MMP systems, though they have somewhat less finely-grained partisan proportionality (limited by district rather than legislative body size.) OTOH, they are still potentially effectively more proportional because the general electorate preference for individual candidates captures more dimensions of both policy and non-policy preference than mere party preference captures.
It’s also possible to have the party lists being ranked by general electorate preference (especially in MMP, rather than full-body party list PR, because in MMP the list is only used for top-up seats which requires smaller, more manageable lists.)
No, you don’t, if it’s a multimember proportional system, like STV. There are some multimember systems your concern applies to, like the “vote for N, top N win” at-large system used for some US city councils, school district boards, and similar bodies: those are simply means to have a multimember body while avoiding either diffuse or geographically-concentrated minorities from being represented. (And the prospect of that particular kind of multimember arrangement being used to neutralize the effect of black voting rights, specifically, is why multimember districts are prohibited by federal law in Congressional apportionment, currently.)
- Schulze Method [0]
- Ranked Pairs [1]
The Schulze method allows for a simple ranked choice ballot, and satisfies more criteria than other RCV methods. Downside is that it is hecking complicated so it can feel like an opaque process. With distrustful electors it's a no go imo.
The Ranked Pairs method satisfies a similar number of criteria as Schulze, and meets a weaker version of later-no-harm. It's also a very intuitive method. The main downside is that the ballots become impossibly long, scaling quadratically with the number of candidates.
EDIT: I was also swapped "later-no-harm" with "independence of irrelevant alternatives". I should stop writing comments before my morning coffee.
Now if only we can change the political zeitgeist from "Ranked Choice Voting" implying Instant Runoff Voting, to RCV implying Ranked Pairs. IRV is merely what victims of the two party duopoly think they want to express their current frustration, but it would fall apart the more we actually got away from the one dimensional spectrum.
As the article mentions, in the real world score voting would just be approval voting where you put a max score on some choices and 0 on others.
And in approval voting you need to think about how others will vote and pick your cutoff point based on who you think has a chance - do you vote "yes" for the center-right party to avoid the hard right party getting in? Or do you vote "no" to help the center-left party beat the center-right party? (swap those directions to personal preference)
RCV isn't perfect, but in all but the smallest elections there's really no practical strategic voting considerations. You just state your true preference order.
Of course, I'll take any of them over FPTP.
I disagree that strategic voting as a downside outweighs the downsides of RCV or FPTP - especially when FPTP itself is susceptible to strategic voting, too. None of the three satisfy the condorcet winner criterion (that is: none of them guarantee the winner would beat every other candidate head-to-head), but it seems less likely / more contrived for score/approval voting to fail it.
> And in approval voting you need to think about how others will vote and pick your cutoff point based on who you think has a chance - do you vote "yes" for the center-right party to avoid the hard right party getting in? Or do you vote "no" to help the center-left party beat the center-right party? (swap those directions to personal preference)
That's why I'd personally go with a simple three-level score vote: "yeah", "meh", or "nah". If people really want to shoot themselves in the foot by ignoring the "meh" option, then so be it, but at least the option is there for people to vote "meh" for candidates that are merely acceptable/tolerable (and reserve "yeah" for ideal candidates and "nah" for unacceptable/intolerable candidates).
To clarify, I never intended that as a defence of FPTP. It's awful and I'll take any of the systems being discussed here over it. It was a statement specifically towards IRV over score/approval.
https://clayshentrup.medium.com/later-no-harm-72c44e145510
it's not even close.
Score voting calls for ambiguous, proven-to-be-culturally-variable ratings that have no natural relationship to preferences (and approval does the same thing in an opposite way, while score voting relies on ambiguous distinctions that don’t have a clear mapping from actual preferences, approval forces voters to rank with artificial ties that don’t have a clear mapping to actual preferences); most metrics simply do not reflect this because they have no mechanism to.
Where these problems don’t apply, as is the case many places outside of secret ballots for elections of public officers, score and approval can be great. E.g., open ballot approval on a choice of a group activity where voting “approve” is a commitment to participate if the selected option wins makes perfect sense, or open-ballot score voting where the given score is a commitment to provide some quantity of resources proportional to the score to the common effort if the item given that score wins. Then the ratings given are tied to concrete differences in preferences.
(On the other hand, choosing between single-winner ballot methods is mostly a waste of time regardless of the abstract merits of the particular system; the most significant difference within democratic systems in quality of government by popular engagement, popular satisfaction with government, and a number of other measures is with the degree of proportionality of the main legislative body, and no single-winner system is going to deliver much of an improvement there, you need a system designed for proportionality, which means multimember districts on some level.)
utter nonsense. the scores are just normalized utilities.
or, if you're being strategic, you're just min-maxing based on expected utility. https://www.rangevoting.org/RVstrat6
> (and approval does the same thing in an opposite way
utter nonsense. approval voting is just a binary 0-1 scale instead of e.g. 0-5. the math in no way changes.
> while score voting relies on ambiguous distinctions that don’t have a clear mapping from actual preferences
utterly false. there is nothing "ambiguous" whatsoever.
> approval forces voters to rank with artificial ties that don’t have a clear mapping to actual preferences
of course it has a mapping to actual preferences. what you're _trying_ to say is that the transform from utilities ("preferences") to scores is LOSSY. the problem is that transforming to ordinal rankings is even MORE lossy, and this is why approval voting still outperforms ranked methods in general, especially with strategic voting taken into account.
we have several pages explaining this common myth to laypersons.
https://medium.com/@clayshentrup/expressiveness-6ef8c034bc65
https://www.rangevoting.org/Expressiveness
> the most significant difference within democratic systems in quality of government by popular engagement, popular satisfaction with government, and a number of other measures is with the degree of proportionality of the main legislative body
there's absolutely no evidence for this. you're making the "proportional representation fallacy". i address this here.
https://clayshentrup.medium.com/the-proportional-representat...
so virtually everything you just said was incorrect. which is why it's good to study this issue and talk to experts before forming opinions.
utterly false.
https://www.rangevoting.org/HonStrat https://www.rangevoting.org/Honesty https://www.rangevoting.org/RVstrat6
> And in approval voting you need to think about how others will vote and pick your cutoff point based on who you think has a chance
1. this is true of _every_ deterministic voting method. it's mathematically proven.
2. even approval voting is extremely accurate. see voter satisfaction efficiency calculations from harvard stats phd jameson quinn, who served with me on the board of the center for election science. https://electionscience.github.io/vse-sim/
> RCV isn't perfect, but in all but the smallest elections there's really no practical strategic voting considerations. You just state your true preference order.
ludicrous. IRV is extremely vulnerable to strategic voting. my aunt who voted biden even tho she preferred warren (to stop trump) would do the same thing with ranking: rank biden 1st to prevent warren from losing to trump.
something similar happened in the first alaska house special election. palin was a spoiler. she caused the democrat mary peltola to win, even tho fellow republican nick begich was preferred to peltola (and to palin) by a large majority. in other words, palins supporters hurt themselves by honestly ranking palin 1st. they would have been strategically better off burying palin and ranking begich 1st.
and the VSE metrics show that even 100% honest IRV performs worse than approval voting with all voters being strategic.
IRV is simply one of the worst voting methods in existence. not to mention being radically overcomplicated and opaque.
i'm sorry that you posted this before having even an inkling about how any of this works.
The problem is that US specifically has a political climate which can be summed up as, "the ends justify the means". While both parties still ritually extoll the importance of democracy and fairness, as soon as it's down to one particular wedge issue, winning is more important than preserving the integrity of the system.
To a large extent, this is because there's basically zero trust between differing factions - it is assumed that your political opponents will squeeze out every advantage they can out of the system when they are in power, no matter how unfair, and so when you are in power, if you don't do the same, you basically hand them victory on policy long-term. Arguably was originally Republicans who triggered this cycle during the Obama presidency, but it's hardly important now - the point is that once trust is broken, it quickly becomes a positive feedback loop.
So, getting back to electoral system, your average voter looks at those proposals to replace FPTP with RCV or whatever, and the first thing they ask is not, "is this more fair?", but rather, "how will this affect the balance of power?". And in any given constituency, the answer is that it will take power away from the current majority (or plurality) and hand it to their opposition. This is both obvious and easy to explain, so that's the usual agitprop angle, and it sells very well.
Any change is also going to disadvantage the two current parties so they will be hostile to it. Which with the hard-to-amend US constitution makes things very hard to change.
Some form of proportional representation or maybe STV in multi-member districts is actually what you'd do. Single member districts with fancy voting systems just gets you a token 3rd party representation.
* better = more accurate and neutral, preferably using representative language from the article
But, oh boy: we have trouble with simple plurality voting. How would you explain the outcome to people if it was implemented broadly?
It makes more sense for the voter to give their favorite candidate a 10, their least favorite a 0, and grade every other candidate on the spectrum between their two extreme picks. So that still seems to be as informative as you could hope for.
However, I submit that there are (at least) three other crucial discussions to have:
- Does my status as an informed voter bias me toward a system for which I have done all the homework?
- Does my script that works fine in my dev system with relatively clean test data scale to an enterprise deployment and real-world loading?
- How significant in improvement is needful to pay for the switching costs? Twice as good? An order of magnitude?
Because I submit that, for all the virtues we informed types can see in RCV, it is prone to malicious attacks, e.g. flooding the ballot with a bunch of relatively unknown candidates.
Furthermore, if there is any sort of contention involved, the complexities are going to lead to (likely unfounded) claims of tinkering.
I'm a huge fan of the ditt-kuh-pow! (DTTCPOW, the Dumbest Thing That Could Possibly Work):
- Make Election Day a federal holiday
- Paper ballots
- In-person voting except in strictly controlled, marginal cases (e.g. deployed military).
Everyone trading security for convenience is misguided, in my opinion. You want to mark a ballot and cast it in secret so that, whatever you have to say to appease the shrieking neighbor, you and your ballot are as physically secure as possible.
Staff the process so that pollbooks are <5k people. We get it done with ~20 people in my county for 13hrs on election day. ~240 precincts in the county. Not cheap, and yes, exhausting to staff. So, what? Staff, equip and support them and elections, can be a positive experience for all involved.
But it's true I don't have a shrieking neighbor or brutal spouse to worry about.
Then, suddenly, the weak security of the postal system will look dodgy.
When, as in the CA case, ballots roll in after the polls close and magically ALWAYS flip the decision the same direction, there will be Sudden Shock and Awe.
It's not security vs convenience it's security vs turnout. Our turnout is embarrassingly low, especially on non-presidential election years, we should be expending voting access to methods that have even less friction. If we can conduct banking via an app or website we can vote via an app or website too. We can start talking about reigning it in once we don't have a third of the country not voting.
You deal with the abusive spouse problem by letting an in-person vote override the mail-in ballot while everyone else gets an easy way to vote.
Rest assured: any workable angle you can think of an many you have not have already been employed to cheat elections.
This is not a partisan dig; rather, simple reality.
I'm not sure MMP is the ultimate ideal, but I much prefer it to FPP. It's a very rare government that isn't a coalition government, and while we get the occasional tail wagging the dog moment, we also get a more diverse set of parties in power, and also more diverse representation.
Often what voters want to know before voting is who a party is willing to go into coalition with, there's usually no big surprises (e.g., the righter-wing ACT party and the centre-right (by NZ standards, they're still left of the Democrats at least when measured by public healthcare etc.) National always dance together, Labour and the Greens typically get along via a confidence and supply agreement even if the Greens are not formally in coalition) but then you have New Zealand First, who are openly mercenary about their choices, and the original Maori Party that went into coalition with National which doomed them in the next election.
[0]: https://en.wikipedia.org/wiki/Mixed-member_proportional_repr...
This acronym needs to be more widely deployed.
Also the problem in some countries isn't the voting system alone, it's the unlimited donations by billionaires, unchecked gerrymandering and general institutional decay.
System doesn't matter when people get away with ignoring it.
The trick is coming up with a system that diverges towards the interest of the masses, rather than the elites or the corrupt or the foreign. Basically, the system must get the voters' preferences wrong, but get them wrong right.
This is really akin to the nature of representative democracy. Elected leaders manage the voters' interest rather than always doing what is polling better on that day. It's obviously a tricky task, but it's the hallmark of the best-functioning democracies.
Right now we're about as far from the goal as we've ever been. Our systems have been hacked and we don't have an effective counter.
The system just needs to be wiser and more resilient. For instance, the Electoral College was originally intended to be more than a bunch of sworn minions. In Europe, many parliamentary democracies have strong party infrastructures that prevent their takeover by "memetic" forces. There are endless possible improvements.
Barring that, RCV would do wonders for the current political ecosystem.
Elections are only good for finding the most popular candidate instead of the most qualified for office. Keep terms medium length (e.g. four years) and you'll have people focus on the greater good instead of personal benefit because at the end of their term they have to live in the society they helped build.
And of course the Legislature had gone in and muddled with them even before they made the ballot (which, to be fair, is explicitly the Legislature's job in the WA initiative system, but it didn't help the confusion).
It was not a pretty election. I am convinced that direct democracy is not the answer, and that was with just four relatively inconsequential initiatives.