[EM] Why I think IRV isn't a serious alternative 2

Abd ul-Rahman Lomax abd at lomaxdesign.com
Mon Dec 29 19:49:42 PST 2008

At 03:48 PM 12/29/2008, Terry Bouricius wrote:
>Abd wrote:
>The term "majority" as applied to elections has some very well-established
>meanings. If we say that a candidate got a majority in an election,
>we mean that a majority of those voting supported that candidate.
>There are quibbles around the edges. What about ballots with marks on
>them but the clerk can't figure out what the marks mean? Robert's
>Rules are clear: that's a vote, part of the basis for a majority.
>I guess a little rehashing is needed to correct Abd's miss-stating of
>Robert's Rules of Order on the basis for determining a majority.

Pot. Kettle. Black. We'll see:

>  Abd seems
>to be relying on RRONR description in chapter XIII on Voting, on page 402
>of how to deal with "illegal votes," such as over-votes, cast by legal
>voters -- they should be included in the denominator for calculating a
>majority. However, on page 387 RRONR states that "majority vote" means
>"more than half of the votes cast by persons legally entitled to vote,

Yes. A "blank" is exactly that, a *blank ballot.* This has been 
referred to as "scrap paper." It is moot, it is not counted for 
anything, though the clerk may make a note of it. "Abstentions" 
generally refers to no vote at all.

Bouricius has acknowledged that illegal votes are nevertheless 
included in the basis for majority. But he will try to turn 
"abstention" into some reference to incomplete ranking. In other 
words, overvotes count, but not exhausted ballots? He's stretching 
desperately for some way to claim that what FairVote has been doing 
for the last decade has been legitimate.

>  The question is
>whether an exhausted ballot (one with no preference shown between the
>finalists) in an IRV election, is an abstention or an "illegal" vote.

Please do remember, as well, to consider the case where the voter has 
filled out all available ranks on a ballot.

That vote is neither an abstention nor is it an illegal vote. It's a 
vote for every ranked candidate. On a Bucklin ballot, all those votes 
would eventually be counted. There is no precedent at all for 
considering these votes as abstentions or illegal. If they are 
considered illegal, the rule would be, in public elections, 
unconstitutional in the U.S. Were the votes for Nader in Florida 2000 
"abstentions."? If the Nader voters believed Nader's argument that 
Gore and Bush were Tweedledum and Tweedledee, and, even on an IRV 
ballot, they only ranked Nader, would that have been an abstention? 
No, it would have been a vote for Nader, not an abstention. They 
voted, they didn't vote for the winner, the winner did not get a 
majority, not matter what tricks you do, what shuffling you do with 
the ballots.

What is found with IRV is a "majority of ballots containing a vote 
for continuing candidates." Not an unqualified "majority.*

>Since RRONR mentions "abstentions" rather than merely using the word
>"blanks," it can be interpreted that there may be some way of indicating
>abstention, other than with a blank ballot.

"Abstentions" are explicit. You have to remember that the statement 
is a general one, it doesn't only refer to written ballots. A member 
abstains informally simply by not acting to vote in any way. And a 
member may abstain formally, and it will be recorded as such under 
some circumstances, by stating -- or writing -- the abstention. The 
intention of one abstaining is to have no part in the process, and 
this must then exclude the abstaining voter not only from the votes 
but also from the basis for majority. This is *entirely* different from voting.

>  I think this perfectly fits
>the concept of an exhausted ballot, where the voter has abstained and
>indicated no preference between remaining candidates, if the voters
>favored candidates cannot win.

This assumes that the "abstention" is a deliberate act, so this 
argument utterly fails with RCV ballots. And this does total violence 
to Robert's Rules. Their meaning is very clear. As I've mentioned, 
Australian law is explicit about this: they use the term "absolute 
majority" to refer to the quota in single-winner STV, mandatory full 
ranking, and they drop that word and use something like the "majority 
of ballots containing a vote for the continuing candidates," with 
Optional Preferential Voting.

In any case, the ballot arguments I cited were quite explicit: "a 
majority of the ballots". There was no mention of excluding any 
ballots at all, and, as you know, Robert's Rules doesn't accept this 
argument about majority. It claims that if voters don't fully rank 
the candidates, "it may prevent any candidate from receiving a 
majority and require the voting to be repeated."

It is impossible to reconcile this with the argument Bouricius is 
making. He and his friends have redefined "majority," and have used 
that redefined word in propaganda promoting IRV, without making the 
redefinition clear. It is a sophisticated form of lying, if done 
deliberately. And I saw what happened when I started to edit the 
Wikipedia article to make the distinction clear: it was vigorously 
resisted by Bouricius and others. I'm afraid that they do not want 
voters to understand this, when they are considering replacing 
majority-required runoff voting with IRV. They are quite happy to 
lead them along.

It may have backfired with Bucklin, you know. Bucklin was sold as a 
method for finding majority winners, which, of course it does 
sometimes, probably a little better than IRV, in fact. But when it 
was realized that it did not always do this (and under some 
circumstances, like IRV, it *rarely* did this), my guess is that the 
complexity of the ballot was no longer considered worthwhile. An RCV 
ballot, from the voter's perspective, is the same as a three-rank RCV 
ballot, except the voter may vote (with Duluth Bucklin) as many 
candidates as desired (only in the third rank, which becomes pure 
Approval). But, of course, the STV method for counting IRV is far more complex.

I'm really appalled that Bouricius is insisting on this, I thought he 
had more integrity than that. It's possible to argue with a straight 
face that there are problems with real runoffs, such as allegedly 
lower turnout, and in order to make my case there I need to introduce 
new concepts, though I'm rather surprised if it is actually new, 
i.e., my observation that low turnout in runoffs may simply reflect a 
normal occurrence: lower preference strength among the voters 
regarding the candidates. And that such voters abstaining may 
actually improve results. *That's* abstention. It is a separate decision.

Is an IRV election a single election? (Legally, it certainly is!) Did 
a voter vote in that election? How can you say that the voter 
"abstained" if the voter actually cast a legal vote? The voter simply 
did not vote for a winner. Does it matter if the voter voted for the 
runner-up or not? Why is a voter who votes for the runner-up a part 
of the basis for this alleged majority but not one who voted for someone else?

The question of whether or not voters can be compelled to rank, in 
the U.S., was quite firmly addressed in Dove v. Oglesby. They cannot. 
And to exclude them from the basis for a majority because of the 
details of IRV procedure is quite clearly unjust. Take the same 
ballots and interpret them with IRV or with Bucklin. Bucklin will 
tell you, perhaps, that no majority was found, so-and-so was the 
winner by a plurality. Same voters who voted for that candidate on an 
IRV ballot, same electorate voting overall, IRV reports a majority.

Difference? Well, IRV doesn't report a majority, it reports something 
more specific: a "last-round majority." It's actually only a 
technical detail of counting, it is not necessarily an ordinary 
majority. I.e., the kind of majority that Robert's Rules and normal 
democratic process require.

No, Robert's Rules of Order is very clear, overall. Ambiguity in how 
they described the actual process allowed you and yours to claim that 
they were describing the IRV process and the IRV "majority," and get 
away with it for a while. (And I think you and many others were 
convinced this was true -- *I was convinced at first, though I 
smelled a rat. Why was RRONR allowing election by a plurality? It was 
some time before I noticed the comment about educating the voters, 
which clearly revealed that they understood what I understood and 
what I'd thought any parliamentarian would understand.) The comment 
about educating voters as to the consequences of not fully ranking 
gives it away. No parliamentarian would have interpreted the 
preferential voting method in RRONR as redefining majority, it is a 
totally well-established and understood concept.

>  There is room here for reasonable people to

Sure. There is, in fact, more: there is room for liars to lie.

The position Bouricius has taken here, that the IRV "last round 
majority," called simply a "majority" by most IRV proponents, but 
"majority of votes" by some or "majority of ballots" by some or even 
"more than 50% of the total votes" in the ballot arguments in Santa 
Clara County in 1998, is what the ordinary person and the informed 
will consider a "majority," is not at all reasonable, and, since I'd 
believed that Mr. Bouricius was reasonable, I must begin to suspect 
something darker.

(Steve Chessin signed this in those ballot arguments: "Measure F does 
NOT repeal the majority requirement for electing county officers. A 
candidate still must earn more than 50% of the total votes cast in 
order to be elected; it just allows the majority winner to be 
determined in one election by using the more modern technique of 
Instant Runoff Voting.)

>  Perhaps an organization could reasonably write bylaws to
>expressly include or exclude such exhausted ballots from the denominator
>in determining a majority threshold.

If Robert's Rules of Order applies, they are included unless 
specifically excluded. Excluding them is the equivalent of allowing 
election by a plurality. IRV is a method which *may* find a majority, 
but which completes with only a plurality of voting members who cast 
a vote for the winner. This was not the intention of RRONR in 
suggesting the possible use of preferential voting; rather, it was to 
be *better* than election by a plurality. That's not "Plurality 
voting," but the generic matter of electing someone when less than a 
majority of those voting in the poll voted for the winner. Thus 
Approval Voting, Range Voting, etc., are plurality methods in this 
meaning, if a majority isn't required.

There are only a few methods for guaranteeing a majority.

(1) Constrain the voters to vote for only one of two possibilities. 
The most common example is Yes/No. If you leave the question blank, 
that's an abstention -- or a blank if there is nothing else on the 
ballot. With elections, restrict the ballot to two candidates and do 
not allow write-in votes.

(2) Require all voters to vote for every candidate but one. This is 
the STV method with mandatory ranking, as used in most of Australia. 
The votes aren't counted simultaneously, but any method which does 
this can guarantee a majority. This also does not work if write-in 
votes are allowed.

Top two runoff with only two candidates on the runoff ballot, then, 
does guarantee a majority if write-ins are prohibited. (However, 
prohibiting write-ins in order to guarantee a majority is putting the 
cart before the horse. We don't find a *true* majority if we have 
constrained the voters from free expression, we have simply forced 
them to make only one of two choices. With elections, there is always 
the third possibility, in theory: no election. Majority failure is a 
form of voting "none of the above." The majority preferred none of 
the above. We can see this clearly if it's an Approval ballot with 
Yes/No voting.)

This clarity and this uncertainty of success is exactly what Robert's 
Rules prefers to see. Majority required, and no guarantee of it. We 
do not fulfill the democratic goals behind this by creating pretense 
about majorities. We may, for expediency, be content with a 
plurality, but we shouldn't pretend that this is democratically 
optimal, or "majority rule," it's a compromise due to practicality. 
RRONR suggests preferential voting, not to guarantee a majority, for 
clearly they know it does not, but to more commonly find one, thus 
avoiding inconvenient runoff elections.

RRONR contemplates that voters will be encouraged to completely rank; 
if they do, there will, indeed, be a majority, and it will be a 
genuine one if the voters were not coerced. But, of course, that 
works with other forms of preferential voting as well, not just IRV.

>  If the organization wrote bylaws to
>include exhausted ballots in the denominator, then an election could fail,
>requiring some alternate procedure (or new election) to fill the office,

Robert's Rules *assumes* this. Election by plurality is prohibited 
unless specifically authorized in the bylaws, which is discouraged. 
They also state that preferential voting must be authorized in the 
bylaws. I'm not quite sure why, in fact. If a preferential or other 
ballot process finds a majority of voters supporting a result, 
clearly accepting it, then I can't imagine why there would be an 
objection to it from the point of view of the rules. There remains 
the possibility of multiple majorities with some methods, and there 
is clear precedent for handling that.

In any case, if an organization were to adopt IRV, stating that the 
method is as described in RRONR, or if they were to simply place the 
method description in the bylaws without adding additional 
assumptions, the method would continue to require a true majority. 
FairVote has snookered a host of organizations into doing something else.

While, at the same time, claiming that this method guarantees 
majorities, or is "majority rule," or with arguments like that.

>or the bylaws could be written to exclude exhausted ballots so that the
>one election would be decisive using a reasonable definition of a
>"majority vote" (using RRONR's standard definition that EXCLUDES
>abstentions in determining a majority threshold.)

Bouricius has totally invented a new meaning to "abstention." It has 
meant, until now, the action of someone who deliberately decided not 
to vote. Now it means, according to him, someone who fails to vote 
for a frontrunner, one of the last two (or sometimes a few) 
candidates left standing in the IRV process. It doesn't matter if 
they voted for someone else, those minor candidates don't count.

It doesn't matter if they used up all the ranks, believing that IRV 
encouraged sincere voting, so why not vote sincerely?

All that matters is that FairVote gets another "success." It's going 
to backfire; I'm hoping that we can head it off, and get some 
sensible reform in place, reform that will actually work.

The mistake was made before with Bucklin. But if, when Bucklin was 
dropped, they had instead kept it as the primary round in a two-round 
system, they'd have had quite an advanced system.

Bucklin is really majority-criterion compliant Approval. Any Bucklin 
election which didn't find a majority was a full-on Approval 
election, so the claim that Approval wasn't used in the U.S. is 
ignorant nonsense. Bucklin also had a Range implementation attempted, 
though I don't think they realized what they were doing: this was 
Oklahoma Bucklin, with fractional votes for lower ranks.

American Preferential Voting was a common name for it. We should have 
kept it, and simply added runoffs when there was majority failure; 
Bucklin probably finds a majority a little more than IRV, in 
nonpartisan elections, because of significant votes for the winner 
hidden under the runner-up, votes that are never counted in IRV.

IRV never counts *most* of the votes, commonly. Bucklin counts every 
vote in the rounds that are reached, and, it seems, it was easy 
enough to count them all that they did, even when the counting was 
moot (i.e, a majority had been found without the round), at least I 
think I've seen that.

In any case, I was genuinely surprised to see Bouricius continue to 
debate this point about majority. This is not a college debate, and 
I'm here not to "win" but to explore and discover and express the 
truth, and to share that. I find arguing beyond all reason to be 
utterly reprehensible; it's relatively harmless here because it will 
be exposed, but, we have to remember, someone like Bouricius has 
credentials. He can stand up and speak to some commission and 
committee and people will think there must be something behind what he says.

I was just reading a presentation by some IRV advocate to a Los 
Angeles commission; the same tired arguments, it doesn't matter how 
many times they have been exposed among experts, these people know 
that the sound bites work, they can deceive people with them (and 
many of them have, I'm sure, deceived themselves as well. People 
*want* to believe in their cause.)

But it backfires. History will repeat, unless we start to do something better.

More information about the Election-Methods mailing list