International support for approval voting?


It isn’t viable currently, but if enough voting reformers get convinced about it, it might have a chance. For voters, Asset makes it more interesting to vote, as picking your honest favorite can actually make a difference. Under Approval, I suspect most voters won’t enjoy having to constantly pick the most palatable moderate, while their favorite has no power.
I’d say that if you show voters Asset creates more choice and honesty, they’ll be happy to try it out. This backroom deal stuff falls apart if Asset gets talked about more publicly, with more crosstalk and input, rather than academically. At least, that’s my guess on how it goes.

Single-winner Asset

I have shared this argument over on where I hope it will draw Mark Frohnmayer’s attention.


If you add Martin Luther King Jr to the list, team good has a 3-way split while team evil has a 2-way split… but of course then you just vote against either Hitler or Messolini.

Fine. Just change your fundamental principles to the logical conjunction of Balance and NoFaveBetrayal.

Under Score Voting, you can give the centrist a 7 or 5 so that they get points, but not as much as your favorite.


Now that our friend Marylander has demolished the idea that balance is sufficient, do you know any good arguments to the effect that balance is even necessary (in the single-winner context)?

Yes and you can do the same thing in Approval. Approve the centrist with a .5 or .7 probability, while certainly approving your favorite and her clones. Actually, I think that .99 is more effective than .5 or .7 for the kinds of ratios of support we see today.


Well, I am just saying, if you want to talk to someone who thinks Balance is a fundamental principle, it is easier to AND it with something than to replace it with something. NFB is enough to include Score, Approval, and STAR but kill off like every ordinal system ever invented*.

*except that one other weird system that no one cares about.


I would go with some sort of independence guarantee. I don’t know if absolute compliance with IIA is necessary (which would exclude STAR, though NFB would too), but you need something strong enough that people don’t avoid running out of fear of being spoilers.


“Approval Voting” has been promoted ineffectively. First of all, we need to understand that democratic voting systems, as used in nongovernmental democratic organizations, universally will make no decision without the explicit approval of a majority of those voting. The abandonment of that principle causes, then, many pathologies.

We often confuse Vote-for-one with Plurality, because that’s the ballot used, and voting regulations routinely disqualify ballots showing more than one vote.

So Approval simply means “Count All the Votes,” and I believe that could be a winning slogan. It really is that simple. Roberts’ Rules has the clerk disqualify votes for more than one, giving the argument that the voters’ intention is not clear. That was a blatantly defective argument, but it’s not much of a problem because that ballot will still be counted in the basis for “majority,:” and a majority is required, not the faux majority of IRV that pretends the “exhausted ballots” were not votes.

IRV would be a greatly improved method if not for rules preventing equal ranking. Asset could allow voting for more than one, and because Asset wastes no votes, it would then be necessary to split multiple approvals.

I have suggested that Score voting use an interpretation of voting range, explicit on the ballot, that defines an approval score (probably midrange), so that majority approval can be determined, and if there is no majority, then the election could become Asset, and if that fails, the election could be repeated, as normal under Roberts’ Rules if no majority is found.
(Thus additional meaning is given to the vote. Voting at the approval level or above means “I prefer the election of this candidate to the election being repeated.” Below that level, “Repeat the election if no candidate is approved by a majority.” Thus the voters decide. Commonly, in assessing voting systems, people assume there is some “sincere rating.” In fact, the best theoretical studies of score show that the method is most effective if voters shift the ratings according to election probabilities, considering them as utilities. In translating some internal rating to votes, voters will sanely consider the value of the vote, and not waste voting power on distinctions not important to them. Providing an actioned meaning to rating level, as suggested here, increases sincere expression, since that preference (for completion vs ideal result) is probably sincere.

Instead of realizing that (1) what we call “plurality voting” is almost approval, requiring only a non-coercive simplification of voting process, and (2) an ideal voting system is probably a hybrid of what have been considered alternative systems, and (3) pathways to reform must be identified and enabled, instead, we argue about the “best method,” often based on speculation and not actual experience.

The widest-used “advanced election system” is runoff voting, which almost respects basic democratic principles, violating them only when the runoff has unnecessary voting restrictions. Those seem “necessary” because the goal is to create a majority, but if this is done by forcing voters to approve one-of-two in a candidate election, what has been done is to prefer appearance to reality. Coerced approval is not approval.

So, to my mind, the simplest election reform is to take existing runoff election rules and (1) Count All the Votes in all elections, (2) Allow write-in votes in all elections (including runoffs).

This would do no harm, and would either have little effect or would reduce the need for runoffs.

So the reform to be advocated depends on the status quo.

1. Plurality. Count All the Votes is the simplest reform, harmless at worst, costing practicaly nothing in voting machines, etc. Obviously fairer than tossing the votes. Allows write-ins a real chance, i.e., one could vote for a write-in and a front runner.

2. Runoff Voting. Count All the Votes and allow write-ins, the same. Runoff Voting with these rules is superior to all other systems without runoffs. Runoff Score has lower Bayesian regret than straight Score, perhaps because the latter can elect with a merely plurality.

3. IRV (RCV). Count all the Votes. Do not elect without a true majority of votes cast, hold a runoff instead. Simply allow write-ins in the runoff, and if the goal of creating a result with no more fuss is maintained, then, first, with the features, a mere plurality result is less likely (it’s quite common with standard IRV, IRV has been deceptively sold), and second, by this time, a plurality result is probably going to be confirmed with a third election anyway. But with Asset, a later reform, probably, that result becomes extremely unlikely.

For further reform, use Candidate proxy or Asset voting. For nearly perfect democracy, create indirect elections by an “electoral college” for full-on proportional elections with no votes wasted, and then elect officers through a proportional assembly, parliamentary system (or possibly by the College, which would then be “single-winner Asset.”)

Keep it simple. Overall slogan: Restore Full Democracy.

(Full democracy is not a strange thing. It’s Roberts’ Rules of Order. The election reforms allow an approach to the known full-democratic processes used in many non-governmental organizations (and, actually, within elected assemblies themselves, i.e., the Speaker of the House, for example, must be approved by a majority of members, not a mere plurality, and they will keep voting until that result is obtained, and they don’t restrict the candidate list.)


Well they can run even more extremist “evil twin” versions of themselves to draw negative votes away from themselves, since you can only negative vote against one candidate. Systems that allow you to express an opinion about every candidate don’t have this problem.


Doesn’t asset voting still suffer from center squeeze? The good Condorcet winner type candidates don’t get enough votes and have to transfer their votes to more extremist candidates.


I doubt it. Most extremists would rather elect a centrist than let their opponent get elected. The minority extremist faction would form a coalition with the centrists to avoid a plurality extremist victory.


So you think a more extremist candidate with 40% of the vote will voluntarily give their votes to a moderate with 10% of the vote?


If the moderate was planning to give their votes to the opposite extremist candidates, then yeah. It’d be the politically safest choice, and from a strategic point of view, it’s the only way to ensure the other side doesn’t win.


I highly doubt that. I think it’s much more realistic that the candidates with the highest votes will not give them up and will insist that others have to give votes to them


The moderate could instead threaten to cast their votes for the plurality extremist, convincing the minority extremists to support the moderate. If neither side compromises, the plurality extremist wins by default, so the minority extremists have nothing to lose by voting for the moderate.


So why not just give every candidate an equal number of votes to negotiate between themselves? It sounds like voters have no power in this system


If every candidate had an equal number of votes, then I’m not sure what would be democratic about it. Voters have power over where their vote initially goes, which is how they make sure that only the best candidates are negotiating.

  1. Voters are still there to give power to the moderates. For example, there may be several moderates. The one that has more votes may be the only one that the minority even CAN enthrone.
  2. In a district with, say, 5 members, the district may split 4-0, 3-1, 1-2, etc immediately before negotiations even come in to play.
  3. Perhaps the “moderates” are lazy and disliked by both sides. Then the extremist with more votes will probably win as the bad “moderates” distribute their votes near equally.


If every candidate had an equal number of votes, then I’m not sure what would be democratic about it.

It would be exactly as democratic as what you described. The votes given to the candidates by the voters are irrelevant, since the candidates will (supposedly) just negotiate those votes until they go to the moderate candidate, who wins.


That’s not what I mean. “Moderate” just means “most similar to the average voter on every political opinion axis”, which includes the laziness axis. A candidate who doesn’t match the typical voter on the laziness (/desired laziness) axis would be an extremist, not a moderate.


No, it wouldn’t. Supposing you had 5 liberal candidates, 1 moderate, and 1 conservative, the liberals would always win in Asset under your idea. It’s only when votes are given to your favorite candidate that you’d see moderates winning. If 30% of the electorate is liberal and 60% is conservative, the winning moderate under realistic Asset Voting would be conservative-leaning, not liberal.