[EM] SEC quickly maximizes total utility in spatial model

[Abd ul-Rahman Lomax]

At 07:28 AM 10/26/2009, Jobst Heitzig wrote:
>Dear folks,
>
>earlier this year Forest and I submitted an article to Social Choice and
>Welfare (http://www.fair-chair.de/some_chance_for_consensus.pdf)
>describing a very simple democratic method to achieve consensus:
>
>
> > Simple Efficient Consensus (SEC):
> > =================================
> >
> > 1. Each voter casts two plurality-style ballots:
> >    A "consensus ballot" which she puts into the "consensus urn",
> >    and a "favourite ballot" put into the "favourites urn".
> >
> > 2. If all ballots in the "consensus urn" have the same option ticked,
> >    that option wins.
> >
> > 3. Otherwise, a ballot drawn at random from the "favourites urn"
> >    decides.

Well, I find it hard to believe how wrong-headed this is. I won't 
achieve consensus, ever, unless the society is so connected that it 
doesn't need to vote!

I've found the logic of consensus inexorable: if we want to maximize 
consensus in a democracy, we must have a society which values 
consensus. I have no doubt that the method described, in its logic, 
is quite clever. Unfortunately, people don't play game theory, that's 
well known; the game-theoretical predictions don't work. In a real 
society that is large enough, the consensus urn will never choose a 
winner unless there is a true consensus process already in operation, 
people will not naturally agree on a large scale, and, while in small 
organization, 100% consensus is attainable, attaining it in very 
large ones is next to impossible. With 100,000 voters, at least one 
of them, even if they all agreed, would accidentally mark the wrong choice.

Are write-in votes allowed?

In any case, I've come to the conclusion that collective 
decision-making must be deliberative except as to one aspect, 
consent. It is traditional in democracies that no collective action 
can be taken without the consent of a majority. Nothing. Election 
under Roberts Rules of Order requires a truer majority of marked 
ballots cast, and the voter can mark on the ballot "None of these 
jerks," and it's a valid ballot even if there is no candidate by that 
name. It counts in the basis for a majority, *as it should*. It's a 
No vote on all the candidates on the ballot. And the voter can, 
similarly, vote for any eligible candidate; normal small-organization 
ballots don't have names printed on them anyway, they are just blank 
pieces of paper. (Hence if you don't mark the ballot, it is 'scrap 
paper' and isn't counted as a ballot, though it may be reported as blank.

So Robert's Rules, unless a bylaw permits otherwise, requires 
repeated balloting until a majority is found. IRV is claimed to 
simulate this, but actually it simulates repeated elimination, which 
is not what RR recommends.

Voting reformers, I suggest, must understand that Runoff Voting is 
the most advanced system that is in actual use, it is much better 
than shallow analysis suggests, and Roberts Rules says why: voters 
may base their votes in subsequent balloting based on the earlier 
results. RR does not allow elimination, period. And, in fact, some 
Runoff Voting implementations allow write-ins. California took this 
away in a recent decision that voting theorists seem to have 
completely overlooked, so little attention is paid to Runoff Voting. 
San Francisco, for the last runoff election, decided to outlaw 
write-in votes. Very bad idea, and probably politically motivated. 
Candidate -- who might actually have won the election -- sued. 
California law requires that write-in votes be allowed in all 
"elections." They decided that a runoff was just an extension of the 
original election, so that write-ins were allowed in the first round 
meant that the law was satisfied, so cities were free to prohibit 
them in the runoff.

A loss for democracy, and very bad analysis. All of us know the big 
flaw of runoff voting, the possible elimination of a compromise 
winner in the first round, which winner would beat all others in 
direct face-offs, even by a landslide. Write-ins make it possible for 
the public to fix the problem. With better election methods, there 
wouldn't be a spoiler risk in that runoff.

But aren't runoffs unnecessary if you have a good method? And there, 
my friends, lies the real problem. The holy grail has been the best 
single-ballot method. The method described is best only if it is used 
by an electorate sufficiently knowledgeable to make the best choices. 
I might point out that such an electorate could do the same with 
plurality. The electorate needs to know, to adjust individual 
preferences to choose the ideal compromise, what everyone else 
prefers. And how does it do that? It does it with a poll. A poll is 
another name for an election, only we tend to think of polls as non-binding.

There is no single-ballot polling method that will choose a winner 
approved by a majority, without coercing voters, that's the bottom 
line. And a winner not approved by a majority may, indeed, be the 
best possible choice, given the restriction that repeated voting is 
impossible. That, however, is a stupid restriction, because it is 
clearly possible to hold a runoff in public elections, because 
communities do it, and consider it worth the expense.

Comrades, to arms! The campaign for instant runoff voting has 
targeted cities using runoff voting because it is vulnerable to 
attack based on the argument that IRV is cheaper (probably a lie) and 
finds majorities without the need for a runoff (definitely a lie.) 
FairVote is actually damaging American democracy, it is not merely 
that this is a less-than-perfect reform. It's harmful.

It can be argued that it is also harmful as a replacement for 
plurality elections; Roberts Rules of Order only suggests 
preferential voting where it is impractical to repeat the election." 
And then, when it describes the counting, it mentions that the winner 
is the first one to attain a majority of the ballots. It does not say 
"the unexhausted ballots." It does not say "all the ballots." The 
book was written by parliamentarians, and "majority of the ballots" 
is defined elsewhere. All the ballots. A true majority of those who cast votes.

Terril Bouricius and Rob Richie just debated this point on a League 
of Women Voters list, I'd been asked to comment. Richie claims that 
my "interpretation" is my own and that obviously many organizations 
disagree with me. No, Robert's Rules is clear, but organizations are 
free to do what they want, they can implement PLurality if they want, 
and if they adopt IRV and allow election by a "last round majority" 
that is exactly what they are doing. If they, on the other hand, pass 
a bylaw that says that the winner shall be determined as described in 
Robert's Rules in the secion on preferential voting, then they have 
not adopted "IRV," as is being sold by FairVote, they have adopted a 
system that may sometimes under some conditions produce a better 
result than plurality, though under real conditions as shown by 
actual election results in the U.S., it produces the same results as 
plurality in nonpartisan elections.

Runoff voting, on the other hand, in one-third of nonpartisan 
elections where a runoff is held, there is a comeback election.

In Long Beach, California, there was a write-in candidate who won the 
runoff. That kind of result would not be possible if Long Beach 
passes the rule that San Francisco passed -- which was made moot by 
RCV's adoption there, where the ballot information pamphlet lied to 
the voters: "A winner will still be required to gain a majority." 
Worst "impartial" ballot question analysis I've ever seen, and the 
opponents of RCV, and they were many, didn't even notice it. That 
shows how little attention has been paid to the value of Runoff Voting.

In any case, there are instructions to the clerks in Robert's Rules 
about the importance of educating voters as to the value of complete 
ranking, because if voters don't completely rank the candidates, it 
is possible that no candidate will gain a majority, "and the election 
will have to be repeated." It is totally explicit, and Rob Richie is 
a bold-faced liar. He made his usual "I don't have time to continue 
this" escape.

While random choice has an appeal, where deliberation is impossible 
and where results over many elections will average out, what if 1% of 
the electorate wants to elect a crazy who will start a nuclear war? 
Could we afford to take a 1% chance of that?

Election process is best when it actually seeks consensus, and that 
requires more than one ballot, on occasion. Range ballots are great, 
but explicit approval should be indicated, which could be done by 
setting midrange as an approval cutoff. And if there is no majority 
approval, there should be a runoff. It could be argued that if there 
is more than one candidate with a majority, there should be a runoff, 
but I won't go there now. In spite of FairVote arguments, multiple 
majorities are very unlikely if Approval were implemented, or Bucklin 
or Range, or hybrids that allow multiple simultaneous approvals.