Fwd: Responding to Rob Richie, IRV advocate
Kathy Dopp
From: Abd ul-Rahman Lomax <a...@lomaxdesign.com>
Date: Wed, Oct 21, 2009 at 8:03 AM
Subject: Re: Responding to Rob Richie, IRV advocate
To: kathy...@gmail.com
At 10:06 AM 10/20/2009, Kathy Dopp wrote:
>
> Rob Richie's latest response to Abd ul's email that I forwarded to the
> League of Women Voters, US list. This email from Richie went to the
> LWV, US list. (wish you'd join the LWV, US and the list Abd ul.)
Well, I'd rather not follow the traffic and be tempted to respond to
more. You asked, Kathy, I'm responding. If anyone has any questions,
they may write me directly, abd (at) lomaxdesign.com.
What I'm writing won't be enough, in my experience. There will need to
be others who will take up the points, and who nail them down and
develop consensus on each point. Mailing lists aren't necessarily good
for that, but they are good for generating analyzed source material. I
write lengthy comments, it's not necessary for everyone to read them,
but if one person does, and comes to understand a bit more, it's worth
it.
There are specific issues being raised by Richie, because they are
important to him, or because he thinks he can overpower with his
carefully developed arguments. His arguments work; and because we can
see, clearly, around the U.S., that he's convincing people using false
arguments, what does that show? Is it something we didn't already
know, that clever argument can convince people of what is contrary to
the evidence?
Then he uses the fact that he has convinced people as further evidence
that his arguments are correct. His politics, ultimately, is the
politics of opposition and battle, of confrontation and "winning." All
he cares about is the vote count in the initiative that will implement
something he can claim is IRV, even if it's not what Robert's Rules
describes, even if it's not what is used in Australia, even if it's
not, as claimed, a simulation of real runoff voting, as it misses the
crucial factor of deliberative consideration of the actual runoff.
How he gets those votes doesn't matter to him. And that's typical of
the campaign managers he'd like us to imitate.
Much of my comment focuses on Richie's behavior. There is a reason for
that. Politically, it might be best for me to just focus on the voting
system issues, but the problem is that there is a farrago of arguments
that have been developed over the last decade, many of which have been
widely accepted as being factual. This is mostly Richie's work, and he
is well-known among voting system experts for slick evasion, for being
able to turn a negative fact about IRV into some positive claim, by
framing it in a way that leads people into misunderstandings. I talk
about Richie because I feel I need to warn people that he's really
good at setting up arguments that will make you think you understand
something that you don't. He deals most effectively with people who
aren't familiar with preferential voting systems, and he does it all
the time. If you are not a voting systems expert, and you aren't
willing to think and dig and check and weigh, but are just looking for
a quick idea about this IRV thing, you are well placed to be deceived
by him, he knows what your preconceptions are, and he manipulates them
very effectively.
I have to take the risk that you will think I'm just as partisan, that
I'm just as capable of making up phony arguments as he is. Actually,
I'm not. In fact, I'd rather die than do that. I may be wrong about
something, but I will never tell it to you because I think it would
make you accept what I want. The ends do not justify the means, and,
in the end, the democratic process works best when people are honest
with each other and explore their differences without polemic, seeking
agreement instead of victory. It's been tried with Richie, it doesn't
work, because his goal is to win. We will see, below, what Richie
thinks we should be doing about voting systems and how he approaches
public discussion. I've been watching him do this for years.
He claims that IRV will reduce negative campaigning and foster
cooperation, but he doesn't exemplify cooperative values. I became
radicalized about Mr. Richie because I was privy to correspondence in
which voting systems experts sincerely attempted to open up channels
of cooperation, and I saw how he contemptuously rejected them as
ivory-tower theorists. Personally, I think it's a good thing if there
are some IRV experiments to see how the system actually behaves under
American conditions, but, then, where is the critical analysis of
those experiments?
There are only rosy reports from FairVote, with the actual behavior of
the system itself not being reviewed. As I've mentioned, IRV in the
U.S., apparently matching the Australian experience with Optional
Preferential Voting, which is similar in some respects, generally
chooses, even when it goes into runoff rounds of transferring votes,
the plurality leader in the first round. The effect is the strongest
with nonpartisan elections. IRV is sold based on recounting stories of
partisan election results, which are quite different, they still
almost always choose the plurality leader, but IRV comeback elections
become more common and exist when a third candidate draws votes
preferentially from only one of the major party candidates, and there
is no other minor party candidate with as much support.
That kind of vote transfer is far less likely to happen with
nonpartisan elections, nonpartisan IRV elections don't show "virtual
comeback" from vote transfers, so the worst place to spend money on
IRV is nonpartisan elections. IRV is also claimed to be a simulation
of top two runoff voting, which then plays on the fairly wide usage of
that method. However, it's very easy to show from actual election data
that the results of TTR and IRV are different, and it's also not
difficult to understand why, and why the TTR results more accurately
reflect the more-informed will of the people, making for results that
are more satisfying to voters. In this nonpartisan environment, IRV =
Plurality, to fairly high accuracy, the result of my study surprised
me. And that is where IRV has been most sold.
To Richie, that doesn't matter. Any implementation is a victory he can
then cite in order to persuade more communities. I urge communities
considering using IRV to take a hard look at what is actually
happening, and realize that partisan and nonpartisan elections are
different, and that IRV doesn't "simulate" top-two runoff, except in a
simplified theory that neglects the very important issue of better
voter information in the runoff, as Robert's Rules notes.
Richie attacks the flaws of other systems quite like an opposing
political candidate. If you become familiar with the arguments, the
reality of the voting systems and their behavior and history -- and
Richie is indeed familiar with this -- you'll see. He never
acknowledges the problems of his "candidate," or the benefits of other
"candidates," he only tells you the up side of what he's championing
and the down side of everything else. He lays it all out, in fact, if
one is paying attention. Those who don't pay attention are doomed to
be deceived easily; no blame, in fact, on them, because we can't pay
attention to everything. But if any of you end up on a committee
considering real proposals, pay attention! Dig! Check everything you
are told, and carefully. That includes everything I tell you, if it's
important.
I've mentioned how, even though I knew a great deal about
parliamentary procedure and was shocked to see Robert's Rules of Order
recommending "IRV," because I knew how important finding a majority
was to Robert's Rules, I believed what FairVote had said in their
framing of the excerpt they quote. Many others have, most with less
knowledge, done the same, it is not surprising. But once the
difference is noticed, the FairVote interpretation is blatantly false,
and citing a whole series of people who have made the same mistake, if
they have, proves nothing. Most of those people, if given the
opportunity to notice the easily overlooked sentence in Robert's
Rules, would agree that what Robert's Rules recommends, far from being
the complete runoff-preventing method Richie is promoting based on
cost savings, instead requires a new election if no candidate gains an
actual majority.
This is the paradox: if you want to save cost, and these are
nonpartisan elections, don't use IRV. Plurality produces the same
results for much lower cost, and if you want reform, there are much
better options that are a little more likely to find a true majority
of voters approving a result, and that can be counted for cost about
as low as for plurality.
Richie is not about to admit this, though, because he really wants
that feather in his cap, the recommendation of Robert's Rules. He's
simply acting like a campaign manager. That's how he wants the rest of
us to act, as well. No thanks.
Democracy fails when we fail to deliberate, to carefully investigate
before voting on a decision. To my mind, then, if we want to shore up
democracy, we will attempt, in every aspect of the political systems,
to foster deliberation, the careful examination of fact and argument
in detail. We can't all do that with every issue, so we need to
cooperate with each other, so that committees of people are formed,
dedicated to vetting evidence and arguments regardless of what the
evidence and arguments might imply. No fact is false because it seems
to imply we should do something we dislike, and no fact is true
because it seems to support what we want. To a political activist like
Richie, this is naive nonsense.
With the internet, this vetting process can become very broad, so that
no relevant evidence or arguments are overlooked. Decisions are still
necessarily made by small groups or individuals, but, hopefully, based
on broadly-supported information and argument.
Now, to Richie's response.
> 2b. Re: Fwd: Responding to Rob Richie, IRV advocate
> Posted by: "Rob Richie" r...@fairvote.org cvdusa
> Date: Mon Oct 19, 2009 8:26 pm ((PDT))
>
> Abd Lomax has a lot to say. My blogs at www.fairvote.org/blog help
> explain why I take a different view. Briefly I will add:
Richie avoids noting what he agrees with and what he disagrees with.
As a master at political polemic, he tries to form overall impressions
and shies away from specifics, if he can. He can paint lipstick on a
pig, and does, frequently. Clear evidence has been presented that he's
repeating, after he should know better, false information, so he will
frame that as being simply a matter of having a "different view." He
wants the reader to think that, really, this is just a matter of
different points of view, with him being for IRV and me being against
it. That way, he can fit this into the standard partisan exchange,
it's just about which horse we have chosen to back, and both of us --
he wants you to think -- are simply cheering a different horse. Truth
doesn't matter, it's just about sides.
To my mind, Richie represents much of what is wrong with much American
politics. In a way, it's not his fault, he is just playing the game
that very many play. However, in the end, it's a game that can't take
us into a brighter political future, it will doom us to repeating the
past, until we are able, as a society, to transcend it. That's why I'm
writing this. The specific decision being considered, which is most
often, Top Two Runoff or IRV, actually matters little to me
personally, because I don't see it as critical. So what if a slightly
worse result happens in one election out of ten, at most? So what if
some jurisdiction spends an unnecessary million dollars on new voting
equipment and software to handle IRV? It's not my money, usually -- if
it's my jurisdiction I might feel a little differently! -- and so much
is already wasted that it's a drop in the bucket.
But I believe that when I know something and it might help my
neighbors, I have an obligation to disclose it, not to control them or
tell them what to do, but to help them make better-informed decisions.
And that's what I'm doing, I'm fulfilling my personal
responsibilities, resulting from the accidents of my life that led me
to study voting systems, to study the actual performance of IRV in
U.S. elections and elsewhere, and, as well, to study the arguments
that FairVote has presented in many fora, over the years.
> 1.) FairVote and I didn't come up with instant runoff voting. We
> weren't fighting against Bucklin voting when it was repealed
> everywhere it was tried. We didn't do anything as Dartmouth's alumni
> got rid off approval voting. We didn't have any role in the adoption
> of instant runoff voting in Australia and Ireland in the early 20th
> century. Many of the associations and universities using IRV for
> their elections have done so since before FairVote existed.
I don't think anyone has claimed otherwise. One of the techniques of
polemic is to list a series of true statements, then gradually slip in
what one wants people to accept that might not be quite so true.... I
was going to accept those statements, but then realized that certain
ideas are embedded in them that can be misleading. This might seem
like nit-picking, but each one of these points does, in fact, have
some importance when we get into details of the issues.
1. Rob Richie apparently received the suggestion of a friend to call
preferential voting "Instant runoff voting," and there was obvious
political value to it. The term creates the immediate impression that
it's a simulation of the relatively familiar runoff voting, only
"instant," i.e., convenient. That's only partially true, and the
differences between the real thing and the "instant" version are
substantial. So, true, he didn't come up with the rough method, what
he came up with was the promotional name and pushing it as an instant
form of runoff voting, and that concept is quite new. (I seem to
recall that some writer described preferential voting way back, before
1980, as "instant runoff," but it didn't catch on.)
2. Bucklin voting, with one exception, was repealed everywhere it was
tried, which, to my knowledge, was only political elections.
Single-winner STV (i.e., IRV) was likewise, except for some
non-political applications, to my knowledge. Both were called
"preferential voting." The ballot is basically the same between
Bucklin and IRV, particularly the three rank form often known as
Ranked Choice Voting in the U.S., i.e., as with San Francisco. But the
counting method is different; bottom line, Bucklin counts all the
votes up to whatever rank is necessary to find a majority, whereas IRV
eliminates candidates who are in last place with the votes counted so
far, and then transfers votes of the eliminated candidates. IRV is a
far, far more complex counting method, and the candidate elimination,
as Robert's Rules notes, can cause the best compromise candidate to be
passed over, because the candidate is eliminated before the lower rank
votes are revealed.
What was the exception with Bucklin? Well, two of the Bucklin
eliminations were ruled unconstitutional. In Oklahoma, there was a
Bucklin variant that was problematic, the basic Bucklin method was not
rejected, but a unique detail (fractional lower-rank votes), yet,
because of that, they tossed the whole law. But in Minnesota, the
State Supreme Court ruled Bucklin unconstitutional, following
arguments that were accepted nowhere else. Bucklin was working --
there are documented elections with vote counts, and you can see how
it eliminated the spoiler effect -- and it was very popular, there was
apparently substantial protest against the court decision, but the
court was adamant that a constitutional amendment was required and
rejected it again in a rehearing. What happened elsewhere? I have no
idea, and I've been unable to find out with any internet research, I'd
have to go to the law libraries and newspaper archives, big job. What
I know is that sometime at or before the peak, there were almost a
hundred places using Bucklin, and the political journals were quite
excited about "American Preferential Voting," as they called it,
because it was invented here. While it can be said that single-winner
STV, or IRV, was invented here, really the method is STV, simply
applied with a quota of a majority. No early applications existed in
the U.S. for IRV, it was not until later that it was tried here.
3. I have no idea why the Dartmouth alumni dropped Approval. I'm aware
of at least one academic, also an alumnus, who was very much against
dropping it, but you know how politics goes. A lot depends on details
with these things, and often there are additional political
considerations. People often try to implement the voting system that
will favor the results they want in the next election, and certainly
that explains both the adoption of IRV in Ann Arbor, Michigan, and its
repeal at a special election. It was used in only one election.
4. Sure. FairVote didn't invent "IRV," they merely gave it a
promotional name by which it is known in the U.S., nowhere else. I'm
not familiar with the age of the existing implementations, in some
cases; but Richie -- he didn't today -- often plays up the American
Political Science Association, APSA. This is a really good example of
deceptive polemic. It appears that APSA has never held an IRV
election. It's there as a theoretical option, for use in a contingency
that there are three candidates on a mail ballot. It may have been in
the bylaws for almost a hundred years, it would be interesting to know
when and how it was adopted. Because the real way that the APSA
president is almost always chosen, which is through being picked by a
nominating committee chosen by the current President, will work
efficiently in an organization which is mostly collaborative and where
strong political and divisive opposition might be avoided, there is no
need for them to examine the bylaw. In the study of voting systems, we
can define an ideal winner of an election; I won't do that, it's
complex, but suffice it to say that how voting systems work depends a
great deal in the nature of the community electing, on the basis for
voting decisions, and other factors.
In an nonpartisan environment, IRV and Plurality give *almost* the
same results, exceptions are rare. Better voting systems, using a real
runoff, might improve the result in less than ten percent of
elections. If I make some rough estimates of the frequency of a
three-candidate election in APSA, I'd have to guess, rather wildly,
based on the last mail ballot (two candidates), thirty years ago, at
one election in 900. So fixing the IRV process to make it better might
improve results once in 9000 years. Now, how much energy should APSA
put into improving their voting system?
What is in the APSA bylaws is completely irrelevant, that APSA is an
association of political scientists says almost nothing about the
method, but we can infer from it that almost a hundred years ago, IRV
was considered an innovative method. It's quite possible that this was
before Bucklin became widely known. But Richie repeats APSA, APSA, as
did Bouricius before him on this list.
Because it looks good, it adds luster, as does the Robert's Rules
deception. Polemic. No concern for real cogency or truth.
> What we have done is propose IRV as a good idea and a lot of people
> agree, including; national political leaders like Barack Obama ,
> Howard Dean and John McCain; many leading political scientists (far
> more than support alternative single winner systems -- see a list we
> assembled early in the decade
> at: http://www.fairvote.org/irv/endorsers.htm )'; and all the city
> councils and charter commissions that have put IRV on the ballot
> where it has passed in the last 7 years.
Yup. And all that means is that FairVote has marshalled a series of
convincing arguments, you have seen some of them here. He's good at
it, and he's been successful, though it should be kept in perspective
that only a tiny, tiny percentage of elections in the U.S. are IRV.
Most of the public implementations are in places where top-two runoff
was being used for nonpartisan elections. This is the place where IRV
is the most useless, and where it is even damaging, as can be shown by
the study of results. These places all had top two runoff because, at
some point, they decided that having a majority of voters support the
winner was worth the expense. And then Richie and FairVote activists
came along and convinced them that they could obtain this goal without
the expense of a second election! Few people have the knowledge to
immediately recognize the implications of different voting systems.
FairVote presents IRV -- and, where he even acknowledges that there
are other options, attacks other systems -- with arguments that are
easy to follow, it seems, and convinces people that this is a great
idea.
There is only one problem: the arguments are misleading, and the
exchange of top two runoff for IRV,with nonpartisan elections, loses
the very purpose of having runoff voting, in actual practice. FairVote
never points out the down side, the fact that what they are selling
does not do what they claim, about majorities, except to a very small
extent, and they certainly are not going to mention that there are
other methods which can accomplish this improvement better, and more
cheaply.
Why did they get stuck on IRV? I've explained it. If they can convince
communities to adopt IRV, those communities are then already set up to
run a multiwinner Single Transferable Vote election. STV suffers from
the same problems as IRV, but because of various factors, the
eliminations generally will cause much less damage. If there were a
multiwinner STV implementation on a ballot where I could vote, or even
where it was practical for me to work for its adoption, I'd go for it,
even though I know of better methods for multiwinner elections.
FairVote descended from a group that was formed to promote
Proportional Representation. It's a noble goal! But they realized,
early on, that a big obstacle was the complex counting method. And the
light bulb went off: if they could promote IRV to fix the spoiler
effect, if they could promote it to "save money" from holding runoff
elections, they could move around the obstacle. The money would
already have been spent, supposedly saving money. (Apparently, in
actual practice, the savings are overblown and may even be missing,
but that is far from my expertise.)
That voting systems experts were horrified was not important to them.
They are political activists, like bulldogs. I think they've made a
strategic error. Election reform has come many times to the U.S., the
wave adopting Bucklin was much larger than anything IRV has approached
yet, and political scientists were enthused. But it all disappeared.
What happened? Well, folks, voting systems aren't the whole story, and
they aren't even the biggest part of it. If I'm correct, IRV was
dropped from party primaries because people weren't using the
additional ranks. (FairVote asserts this about Bucklin, but it may be
more true for IRV. People use the ranks, at first, because of the
novelty. But, in fact, if there are two frontrunners in an election,
most people have no reason in the world to worry about lower rank
votes, and it's unlikely that most of those votes will even be
counted.) After many elections in which voting transfer, as with IRV,
have no effect, but there is this continual irritation of complicated
vote counting, it's not hard to understand why IRV would be dropped.
That's why actual election results should be studied.
Some of the problems that IRV allegedly fixes can indeed be fixed, at
much lower cost. That's why people who are really interested in
election reform should look carefully at all the options, and
especially at how the systems actually work, instead of purely
theoretical expectations. If a reform is too cumbersome, if it adds
complexity that isn't used, it may not last, and the conversion
expense, if there was any, was wasted.
That's why I personally recommend Approval Voting, in the simplest
basic form, as a first reform. For no cost at all, the spoiler effect
that IRV fixes is fixed. The ballot is the same as a plurality ballot,
only the instructions allow voters to vote for more than one if they
choose. (Approval voting is symmetrical, you can vote Plurality if you
want, just for your favorite, or Anti-Plurality, i.e., "Anyone but X,"
or anything in between, and your vote has exactly the obvious and
easily-understood effect. It helps the candidates you vote for and is
effectively a vote against candidates you don't vote for.) The
counting is the same, only the programming or rules that discard
overvoted ballots are dropped. Simpler than Plurality, a little. But
most voters will simply vote as they have always voted, for their
favorite. In a two-party system, that's quite appropriate Approval
strategy for most voters. The only difference is with supporters of
minor candidates who know their candidate has no chance to win; they,
with this change, now have the ability to cast an "alternative vote."
FairVote will tell you that they won't want to do that, because it
might "hurt their favorite." But that argument depends on the favorite
being a major candidate, and also the other candidate they might vote
for. I.e., the voter wants to vote for Bush and Gore? It's
preposterous. Richie doesn't care, because he knows that if he gives
the argument, people will buy it, it sounds right at first. Until you
look carefully.
> 2) For all of what Abd says about Robert's Rules of Order, the fact
> is that dozens upon dozens of private associations read it the same
> way that I read it -- and have based their elections upon it, using
> instant runoff voting in a single round of voting.
We don't know how the parliamentarians in those associations read it,
and Richie isn't telling us how he knows how "they read it." I'll
agree that there are public officials who have, on official voting web
sites, claimed that IRV is "recommended" by Robert's Rules of Order.
FairVote says that on their web site, and these public sites reference
FairVote to support their statement. FairVote has repeated over and
over that RRO recommends IRV. That doesn't make it true, at all, and
the reason I "read it differently" is that, though it is easy to
overlook because of where it is placed in the RRO sections, it's
explicit. What they recommend requires a repeated election if a true
majority is not found.
Since Richie continues to insist on this point, here it is, from the
horse's mouth:
Robert's Rules of Order, Newly Revised, 10th edition, page 413-414.
"When this or any other system of preferential voting is used, the
voting and counting procedure must be precisely established in advance
and should be prescribed in detail in the bylaws of the organization.
The members must be thoroughly instructed as to how to mark the
ballot, and should have sufficient understanding of the counting
process to enable them to have confidence in the method. Sometimes,
for instance, voters decline to indicate a second or other choice,
mistakenly believing that such a course increases the chances of their
first choice. In fact, it may prevent any candidate from receiving a
majority and require the voting to be repeated. The persons selected
as tellers must perform their work with particular care.
"The system of preferential voting just described should not be used
in cases where it is possible to follow the normal procedure of
repeated balloting until one candidate or proposition attains a
majority. Although this type of preferential voting is preferable to
an election by plurality, it affords less freedom of choice than
repeated balloting, because it denies voters the opportunity of basing
their second or lesser choices on the results of earlier ballots, and
because the candidate or proposition in last place is automatically
eliminated and may thus be prevented from becoming a compromise
choice."
So, from the second paragraph, RRO does not "recommend" IRV except in
two ways: where repeated balloting is not "possible," and, in
jurisdictions using runoff voting, clearly at least one repeated
ballot is possible, and this becomes almost exactly RRONR's
recommendation if write-ins are allowed in the runoff (because there
are no actual eliminations, merely what might be called a strong
public recommendation to vote for one of the top two. I'm aware of at
least one public election where the public ignored that recommendation
and elected a candidate as a write-in in a runoff election.) The
standard way of avoiding runoffs is to elect by plurality, and most
organizations which want to complete with one ballot do exactly that,
they specifically allow election by plurality in their bylaws, because
RRO prohibits it by default.
When they say that IRV is better than election by plurality, they are
saying that a majority found through vote transfers is better than a
mere plurality. (They may not have been aware that IRV in nonpartisan
elections almost always matches plurality and frequently fails to find
a majority, that was a discovery I made in my own study of the San
Francisco and other recent elections. I later found that this was a
known effect in Optional Preferential Voting in Australia.)
But, to make the meaning of their statement clear, it's better to find
a majority with vote transfers that to accept a plurality result. But
they are not at all suggesting the accepting of an IRV plurality, and
they are explicit, in the first paragraph, that if that happens
(because of incomplete ranking) the election must be repeated. Quite
simply, they don't accept that it is ever truly impossible to hold
another election, it's merely inconvenient, and requiring a majority
for decisions is fundamental to Robert's Rules.
Do those organizations specify, in their bylaws, election by
plurality? If all they did was write that the method of preferential
voting as described in RRO shall be used, which is what Richie told us
in the previous mail, they did not specify that, because RRO does not
specify that. But RRO suggests that the exact details be specified in
the bylaws. If they specify exact details, and they don't use the RRO
exact language, but then write the rules to allow election by
plurality, they have done an end run around the RRO insistence on
Plurality. RRO wants organizations to very carefully consider any loss
of the majority requirement. But there is another possibility. They
use the exact RRO description, but misinterpret it, there is an easy
misinterpretation.
That RRO description could be faulted because it does not specify the
treatment of exhausted ballots, in terms that make it crystal clear
that they remain valid ballots. Other sections of RRO, however, make
it clear that a ballot with "illegal votes" is to be counted as part
of the basis for a majority. How much more would a ballot with a legal
vote be counted? So there is no problem for a parliamentarian here,
only for someone who isn't familiar with the rules. In the counting
procedure section, the ballots are separated into piles, sorted by the
candidate, among those not already eliminated, showing on each ballot
as being in the highest preference, then the smallest pile is
identified and that candidate eliminated, and the ballots resorted. If
any ballot has had all of its candidates eliminated, "it should not be
placed in any pile, but should be set aside."
And then it says that the sorting process continues "until one pile
contains more than half of the ballots, the result being thereby
determined."
What is "half the ballots?" From standard practice, and from the later
comment about educating voters, it's very clear. It is half of all
ballots that are not blank. (Blank ballots are 'scrap paper'). Want to
vote against every candidate? Write No to All on the ballot, or,
anything, really. Under Robert's Rules, this might prevent an election
from completing, as it should. (It should be understood that in RRO
standard, candidate names aren't generally printed on ballots. You
write the name for all votes. Public elections were originally like
that in the U.S.!)
But someone might interpret "set aside" as meaning "eliminated." It is
very obvious from the entire section, though, that this isn't what is
intended. But if people have in mind the idea of eliminating runoffs,
if that is the whole reason they are implementing IRV, they will be
disposed to overlook this detail. Most people, learning about IRV,
don't even realize that it would fail to elect by a majority, because
they are always shown examples of fully ranked ballots. They don't
know that, under stable conditions, most voters won't fully rank, and
RCV doesn't even allow it, when there are more than four candidates.
> None that we have
> found use instant runoff voting with multiple rounds of voting.
I don't doubt it. But that doesn't mean that Robert's Rules of Order
recommends IRV! It means that eliminating runoff elections was
considered more important than obtaining a majority! No voting method
can guarantee that a majority of voters accepts a winner, unless the
acceptance is coerced, which is exactly what Australia does. There,
you *will* vote for all candidates but one, that is, if there are
three candidates, say, and you detest two of them and would not want
to be responsible for the election of either of them, you have no
problem with plurality, you have no problem with almost all methods,
you really have no problem with IRV. Unless mandatory full ranking is
required. You must choose to vote for one of them over the other, even
if you have no preference at all between them. That's the Australian
system which does, indeed, always elect by an absolute majority
(excepting ties, of course). Where there is Optional Preferential
Voting in Australia, election by plurality shoots up. (And the rules
explicitly allow that, or they'd have to repeat the election!)
> That's because a lot of folks agree with us that voters have the
> right to abstain from choices -- if they don't want to express a
> preference between the final two candidates, they don't have to do so.
The consequence doesn't follow from the condition. They don't avoid
multiple rounds for that reason, it doesn't make sense. Richie is
here, without having mentioned it explicitly, referring to mandatory
full ranking, which is a way of avoiding majority failure (through
coercion), but real runoffs don't require the expression of "a
preference between the final two candidates." Indeed, that's part of
the benefit. You don't have to vote in a runoff election, it's
optional, and if you don't care, it's rational! Richie glosses over
the fact that if "runoff elections" are being run under rules that
allow write-in votes, or which have no names on the ballot at all,
voters aren't limited to two candidates at any stage.
What it boils down to is how much effort we are willing to put into
securing democratic results, approved by a majority, or the discussion
and deliberation continues until a majority is found. Democracy takes
work! Repeated balloting, indeed, as Robert's Rules actually
recommends, is a very advanced voting system, in terms of the quality
of results. It's "Condorcet-compliant," for starters, assuming the
lack of restrictive rules. One of the problems with voting systems
theory is that it has focused almost entirely on finding ideal methods
that always produce a result, a "social ordering," in a single ballot.
Real democracy, deliberative process, requiring a majority, probably
does better than any single-ballot system, by allowing iterative
consideration. Good single-ballot systems are important, but are not
the true nuts and bolts of democratic process. Rules of order
generally attempt to confine all votes to Yes/No questions, with
amendment process that is also Yes/No, and thus the known paradoxes of
single-ballot voting systems don't apply to real deliberative voting,
they only arise when there are three or more choices to be chosen from
at once.
> 3) I'm paid, to be sure, but that's not why I do this work -- indeed
> I worked for FairVote for a long time without pay. And nearly
> everyone out there working for instant runoff voting today is not
> paid. Most of the IRV ballot measures have been run by volunteers,
> for example, and every single one certainly has depended on
> volunteers. Abd and Kathy (if she really wants something other than
> plurality voting) need to get out there, engage with policymakers and
> see what it's like trying to explain the value of an alternative voting system.
I'm sure it's hard, and I can tell you I'd have a devil's time trying
to convince policymakers to implement IRV. Once upon a time, I thought
it was at least a reasonable reform, but I'm now convinced that the
improvement over plurality is too small to be worth the expense, and
that it's a step backward from top two runoff, not to mention TTR with
write-ins allowed, which is theoretically spectacular, it has not been
given proper credit by voting systems experts, because it isn't
single-ballot, and single-ballot is where almost all the study went.
Absolutely, I'd like to see better than plurality voting. Simple
slogan: Count All the Votes. Just count all the votes! Do you have any
idea how much of an improvement this would be? If a voter votes for a
candidate, count it, no matter what else the voter has done. So
simple. And so powerful. But do I think I can convince "policymakers"
to do this? Not unless they seek the advice! I might be able to
convince a committee that has sufficiently open process and that
iterates its work, coming up with a preliminary report on which there
is then solicit comment, and back and forth until there is substantial
consensus (and minority reports where appropriate). I can also be
fairly convincing in person, under the right conditions. As to
writing, most people don't want to read enough; in order to become
convincing with the brief polemic that passes for political debate,
I'd have to join Richie and develop the quick sound bites that
convince regardless of the truth. Not my cup of tea. Someone else can
do that. I invented the slogan Count All the Votes, for approval
voting. My contribution, and it isn't misleading, it's quite accurate.
I know and understand much more powerful reforms, and it's hard enough
to apply them in small voluntary organizations. Short of those
reforms, fixing voting systems is a band-aid. Bear with me for a
moment.
1. It's possible to have a practically ideal form of representative
democracy, without what we think of as elections.
2. If secret ballot is needed (let's grant that for large public
elections), the ballot is vote-for-one. No printed names on the
ballot, rather, there is a booklet at the polling place with all
"candidates" listed who have paid a small registration fee, to cover
printing costs. You write in or mark the number of the candidate. Any
registered voter may register as a candidate.
3. The votes are counted and become the "property" of the candidate
who received them. I will now call any candidate who received any
votes an "elector." An elector is a public voter. (Kind of like the
electoral college, if it was assigned proportionally.)
4. If this is being used to elect maximum N seats of an assembly, and
V votes were cast in the election, V/N is the "quota."
5. Any elector with a quota of votes gains a seat (if he or she wants
it), and still may have votes left over after the quota is subtracted.
6. Electors may transfer their votes to other electors, or to anyone
qualified under the rules to receive votes.
7. If any votes are not reassigned to create a seat, there will be at
least one seat vacant, until such time as electors find a suitable and
acceptable compromise.
Now, one further step:
8. Seats have the right to participate in the deliberative process of
the Assembly, and to vote on all issues before the Assembly, as would
be expected. However, means will be provided for electors who are
either in attendance at an assembly, or able to follow assembly
process remotely, to vote on all issues, as well. When an elector
actually votes on an issue, the vote of the seat is fractionally
devalued in analysis, reduced by 1/Q for each vote that the elector
"owned." (For the purpose of voting only, it is as if, by voting, the
elector has retracted the vote assignment to a seat, for that vote
only).
Without the last provision, it's a very accurate form of proportional
representation, first proposed by Lewis Carroll in 1884. With the last
provision, it's hybrid representative/direct democracy, with the best
features of both. My opinion is that only rarely would there be enough
direct votes by electors not holding a seat to change a result, but
what is much more important about this proposed system is that the
electors are intermediaries between the general public and the members
of the assembly. Voters will know whom they voted for. And they will
be able to see where that vote went. Any voter who wants to put in the
time and take the risk of being publicly identified can become an
elector, all the person has to do is vote for himself or herself.
Voters have only one decision to make: of all the possible candidates,
whom do they most trust to exercise their vote for them? Don't trust
anyone? Become an elector, though probably it's not going to be worth
the effort. If you don't trust anyone, you probably won't be trusted
yourself, and one fractional vote in the Assembly isn't worth the
trouble!
Okay, One More Utopian Plan, right? Except that this kind of system
could be implemented, and there are less formal versions, in any
organization. I haven't explained all aspects, but I see overall
political reform as coming through voluntary, ad-hoc implementation of
organizational reforms in small organizations. We keep trying to fix
the biggest organization, the government. But if we can't fix small
organizations and make them maximally 'intelligent,' i.e., gathering
the best judgment from the members, and fostering the most powerful
unity and cooperation, how would we think we could fix the government?
Start small. When you see the opportunity.
Let me know how you do! (And if you have questions about any of this,
ask me by email, I'm not following the lists and I only respond to
stuff like this when asked to do so, at this point.)
Back to Mr. Richie.
> 4) If you've ever worked on a candidate's campaign or ever deeply
> cared about a candidate's success in an election, you'll see why
> violating later-no-harm is such a huge problem. Having your ballot
> count for more than one candidate at the same time (as happens with
> range voting and approval voting) is messy. Try to imagine being the
> campaign manager for one of the major party presidential candidates
> early on in their nomination races last year and what it would be
> like with different systems. I suspect you'll quickly get to the
> point where I get, which is that I want instant runoff voting.
My point, actually. Mr. Richie wants me to look at things from the
perspective of a campaign manager, who works for a result that is
fixed and in opposition to the work of all other campaign managers.
Campaign managers hate Later No Harm violation, and they imagine that
the public will hate it too. They may even be right, because the
existing system so much makes us think of winners and losers, instead
of seeking and finding maximally unifying solutions. It's not a
zero-sum game, in fact, and treating it that way causes great
inefficiency -- wasted effort -- and damage.
Bucklin starts out with vote-for-one, as it was implemented. Your vote
isn't going to count for an opponent of your favorite unless you
favorite is not going to win in the first round of counting. That
Later No Harm violation only shows up in subsequent rounds, and you
have to look at whom it affects, not just as some abstract overall
principle, without regard for the nature of the actual situation.
The difference between IRV and Bucklin is that with IRV, your second
rank vote won't be counted unless your candidate is *eliminated.* With
Bucklin it won't be counted unless your candidate isn't going to get a
majority straight off. Does this situation hurt your favorite. Maybe.
But it also might help your favorite. Would you prefer your favorite
to be eliminated (IRV) or to have a chance to gain second rank votes
from other candidates (Bucklin). The only way we can compromise is if
we are going to die otherwise? Yeah, no wonder a political activist in
the Winning is Everything Game dislikes Later No Harm violation!
Is your candidate a frontrunner? In nearly all elections, there are
only two frontrunners, call them A and B. By definition, most people
prefer A or B. The election is almost certainly between A and B, and
C, a third candidate, is just running to test the water and perhaps
make a point. In Plurality, C will get fewer votes than C might
deserve, because of people's reluctance to waste their vote, or,
perhaps, C does get the votes, but then the votes are "wasted," at
least as far as affecting the result is concerned.
(Asset Voting, described above as the Utopian System, is the only
system proposed that is not guaranteed to waste votes, because, with
Asset, there are no "losers," just a process of deliberative
compromise on who will be voluntarily chosen to sit in an Assembly,
and that compromise is minimal and among congenial electors, not the
kind of compromise that is choosing the least of two evils. STV, for
many-member proportional representation, can get close to this, but
without the side benefits.)
For genuine supporters of A and B, there is no question of Later No
Harm, they have no incentive to add another vote at all. If they like
C and want to express some support for C, sure, they are risking that
C might beat their favorite, but surely they won't add that vote if
the risk concerns them, if the strength of their preference for A or B
is strong over C. It's only if it is weak that they will even think of
it.
So the only voters who will be considering, much, adding a second rank
vote, are the supporters of C. They have, with Bucklin, already
expressed their preference, C's first rank vote will quite clearly
show sincere first preference, there is no reason to do otherwise. So
that purpose is satisfied, something which doesn't happen with
Plurality, or, for that matter, with straight Approval. Do they add a
vote for A or B?
Why not? This is the risk: they were wrong about C, C might actually
win. But by voting for A or B, they might put A or B over the majority
mark, in the second round. Or, something which traditional voting
system analysis often misses, suppose C might make it into second
place, and the rules require a majority. If they vote for A, say, A
gains a majority, if not, there is a runoff between A and C and C
might win, because of the new credibility. Yes, if there is a majority
requirement in the first round, there is less incentive to add extra
ranks, under some conditions, because of Later No Harm. But those are
actually unusual conditions, and what's the harm of not adding
additional ranks? By the terms of the problem, there might be a
runoff. If the voter is willing to do the work, prefers to see a
runoff, that is, in fact, the voters, collectively, by their behavior,
deciding that they would rather have a runoff than add another
preference.
There is no harm to the democratic process at all. Richie is focused
on the thinking of those who want a personal victory, and to whom that
matters more than finding the most widely-acceptable solution, the
office-holder who has the broadest support. For myself, I want that
most-approved candidate to win, and I prefer to see that, in fact,
over my personal preference. I don't like losing my preference, but I
very much dislike gaining my preference at the expense of the
majority.
And see what IRV does by satisfying Later No Harm: not method which
satisfies Later-No-Harm can maximize acceptance, because Later No Harm
requires that a candidate be eliminated before lower preference votes
can be considered. The same process that guarantees that the voter's
secondary vote can't "hurt" the favorite, because the lower ranked
vote isn't counted until the favorite is eliminated, also guarantees
that other votes that might even allow the favorite to win aren't ever
counted, because the favorite is eliminated before they are revealed.
IRV prevents "harm" to the candidate by eliminating the candidate as a
matter of process, i.e., in spite of the voter's vote. The method
harms the candidate, not the voter. Cool, eh?
FairVote has claimed that Bucklin was eliminated because of Later No
Harm. There is only the thinnest basis for that; the Minnesota Supreme
Court did cite Later No Harm violation in its decision to declare
Bucklin unconstitutional. However, in their review of the case, they
made it very clear that the basis for their decision was *any kind of
alternative vote,* not merely the Bucklin form. It was one very quirky
decision, one not repeated anywhere. Bucklin was not eliminated
because of LNH fears, and voters mostly don't think like campaign
managers. Voters can decide which is more important to them, the
possibility of their second rank vote creating a win for their second
favorite, or the possibility of majority failure or maybe the win of
the less-preferred of the alternatives. And I prefer for voters to be
making that decision, rather than voting system designers. FairVote
says Thou Shalt Not Hurt Your Favorite, it's Bad.
My strong suspicion is that Bucklin was eliminated because it created
political risk for powerful parties. At that time, there were
extensive minor parties that were moving up in popularity. That trend
stopped. Why? Well, Plurality voting is quite hard on minor parties.
Those who wanted to stop the minor party movement would be opposed to
advanced voting systems. It's pretty clear. So, before it spread to
legislative elections and the like, they stopped it. They had the
power to do so. It's an old story.
Any community which is already using top two runoff could cheaply
lower the incidence of runoffs by using Count All the Votes, or
Approval (the only expense is voter education, all voting equipment
can already handle approval voting). They could lower it even further,
with minimal expense with Bucklin and the counting is much simpler,
again, just count all the votes, only there are more votes to count,
probably. It's still just sum-of-votes in each round, and it is
precinct summable.
(IRV is not, IRV vote-counting is extraordinarily complex, that's why
Robert's Rules mentions the caution of the clerks, and one mistake can
ripple through the system, requiring recounting in every precinct,
since the votes reported from each precinct depend on the first round
results from all precincts. That, I believe, is why Kathy Dopp is
opposed, it's an election security issue, and it is, indeed,
significant. No such issue with Approval, no such issue with Bucklin.)
The best single-ballot system, probably, my opinion, would be a hybrid
between Bucklin and Range Voting, and it's quite possible to get there
step-by-step, one small and cheap improvement at a time.
A very nice system, that would fit with existing top two runoff
systems, would be to use Bucklin for both rounds. Bucklin, because it
counts all the votes can find votes that IRV would leave concealed. In
partisan elections, this is of little consequence, but with
nonpartisan elections and two frontrunners, A and B, there will be
supporters of A who have voted for B in second rank, some, and
supporters of B who have voted for A. This is rare indeed for partisan
elections, but with nonpartisan ones, and with sincere voters who
don't care much about "strategy" -- which may be most voters, actually
-- there may be very many such votes. (I've studied this with the San
Francisco data, and cross-ranking between frontrunners was more common
than I'd have expected, and it's reasonable to conclude that some of
these elections -- most San Franscico runoffs don't find a majority --
would indeed have found a majority.)
So using Bucklin would avoid some runoffs, maybe as many as half,
which is much better than IRV's apparent performance as to finding a
majority. Then, in the runoff, allowing write-ins (California's
default, for example) can safely be done without creating a
vote-splitting effect. At this point, there are two clear major
candidates, clearly the top two as to favorite. If you favor a third
candidate, your optimal strategy is clear: vote for your candidate
first rank and the preferred top two candidate in second rank. If you
don't have a preference, fine, don't add another rank. There may be
majority failure from that, but at this point, the general conclusion
is probably that it's not worth another runoff. The win goes to the
most widely approved candidate, even short of a true majority.
(There is another wilder possibility: Use Asset, i.e., a
vote-splitting candidate can essentially decide the election. But I
can easily see this as being quite unpopular, knee-jerk. Bottom line,
really, is that I don't like officer elections at all, where officers
have fixed terms. I prefer good proportional representation, and Asset
would be practically free of partisan politics, except for the
intrinsic partisan politics of the population, and then elect officers
deliberatively, essentially making a hiring decision. No term of
office, they can be voted out at any time. They are servants of the
public, not temporary kings. This idea is quite common, it's the
parliamentary system, probably more advanced than the U.S. general
system.)
The eliminations of IRV are basically contrary to fundamental
democratic principles, as shown by quite reasonable scenarios. A clear
one is if there should happen to be three candidates who are about
equally balanced as to first preference, but one of them is a centrist
and is the first or second choice of *everyone*. If people voted
according to pure political position, this would be quite possible, it
is rational and it makes sense. However, the centrist has slightly
less first-choice support than the other two, who I will call the
extremists.
In this three-candidate situation, the centrist is eliminated in the
first round, but the extremist elected by IRV would lose by almost two
to one in a direct face-off with the centrist. This is what Robert's
Rules of Order is talking about as the problem with "the form of
preferential voting" they have described, and if there is majority
failure, it might be a good thing that there needs to be an open
runoff! Thais drastic a Center Squeeze result would be unusual,
because the triple balance in first preference is unusual, especially
in a two-party system. But lesser failures certainly do occur, where a
candidate who wins under IRV would lose in a direct face-off with one
of the other candidates. This is proven by the experience of runoff
voting, and all voting systems experts know this, and this is why
practically none of them support IRV except a very few who support it
because of "momentum," and possibly the long-term possibilities with
Proportional Representation.
American Preferential Voting, Bucklin, is much better, it does not
have the Center Squeeze problem. Richie wants you to focus on Later No
Harm, but in order to fix the problem with failing to find a
compromise candidate, which is truly serious, it could be a disaster
for a society under some conditions, LNH must be violated, there is no
other way with a single ballot. How could you possibly identify who
the compromise candidate is, or even if there is one, if you can only
see the first preference votes?
Lipstick on a pig. I told you. LNH is a supposedly desirable voting
system quality that most experts consider positively undesirable.
Richie mixes it with some nice color to make it look good and pastes
it on the face of IRV, turning a truly and dangerously negative
quality into an argument for the system. I tell you, he's really
skilled at this, he's come up with dozens of such arguments, he only
trotted out a few today.
--
Kathy Dopp
Town of Colonie, NY 12304
phone 518-952-4030
cell 518-505-0220
http://utahcountvotes.org
http://electionmathematics.org
http://kathydopp.com/serendipity/
Realities Mar Instant Runoff Voting - 18 Flaws and 4 Benefits
http://electionmathematics.org/ucvAnalysis/US/RCV-IRV/InstantRunoffVotingFlaws.pdf
Voters Have Reason to Worry
http://utahcountvotes.org/UT/UtahCountVotes-ThadHall-Response.pdf
Checking election outcome accuracy --- Post-election audit sampling
http://electionmathematics.org/em-audits/US/PEAuditSamplingMethods.pdf
Rob Richie
Rob
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
"Respect for Every Vote and Every Voice"
Rob Richie
Executive Director
FairVote
6930 Carroll Avenue, Suite 610
Takoma Park, MD 20912
www.fairvote.org r...@fairvote.org
(301) 270-4616
Please support FairVote through action and donations -- see
http://fairvote.org/donate. For federal employees, please consider
a gift to us through the Combined Federal Campaign (FairVote's
CFC number is 10132.) Thank you!
Kathy Dopp
From: Abd ul-Rahman Lomax <a...@lomaxdesign.com>
Date: Thu, Oct 22, 2009 at 12:27 PM
Subject: Re: Responding to Rob Richie, IRV advocate
> Date: Wed, Oct 21 2009 3:23 pm
> From: Rob Richie
>
>
> I've made it clear why I and many other people (from leading scholars
> to grassroots activists) sincerely support instant runoff voting for
> single-winner races add won't repeat those arguments. I also have
> tried to explain clearly why I think having your ballot count for
> more than one candidate at a time in a race for one seat creates
> problems and makes Abd Lomax's preferred systems problematic. My
> silence on the rest of this message should not suggest acquiescence.
This is a 100% predictable Richie response. He's correct, silence does
not suggest acquiescence. It is not necessary to counter every stupid
comment; anyone who thinks so is soon completely overwhelmed, for the
internet is full of such.
Briefly, I suggest noticing that Richie's comment about IRV is that of
a salesman, it's typical sales hype. "Leading scholars." That's the
kind of phrasing you'll see in ads for products of quite variable
quality. Richie is a professional salesman, so this is perfectly
normal. But it's important to distinguish, in considering advice,
whether or not the advice has an agenda behind it other than neutrally
informing you so that you can make intelligent decisions.
Absolutely, many people sincerely support instant runoff voting. Most
scholars of voting systems don't support it as a decent voting method,
but some believe Richie's argument -- he argues this in the rare
instances that he's actually writing to communities of experts -- that
IRV is the only practically possible reform, the only one with
"momentum." For some reason, voting systems experts have neglected the
study of runoff voting, and the implications of differential turnout
have been overlooked. Some experts will, indeed, support IRV, but over
Plurality in partisan elections.
If you are a supporter of a major political party, and you dislike
being damaged by the activities of minor parties, you will be quite
well-disposed to IRV, and it will really help you when these
activities hurt your party more than the opposing party. So the
support of many "grassroots activists" isn't surprising. But some of
them are deluded. If they support a minor party, IRV is likely to
eventually demolish their party, that's the usual effect in the long
term, because it makes it very difficult for third parties to actually
win elections, and if they do manage to build enough strength to
seriously contend for the office, that is when IRV seriously breaks
down. There are other methods which gently and evenly treat an
emergent party, allowing its true strength to be shown without serious
damage to results. One of these methods saw very wide usage in the
United States, with roughly a hundred implementations, far more than
IRV at present, and it should not be assumed that it disappeared
because it was defective. It is more likely it disappeared because it
was working, and those whose careers depend on flaws in Plurality had
sufficient power to eliminate the danger before it spread to elections
that affected them.
Then, as to the only argument of substance above, about "having your
ballot count for more than one candidate at a time," Richie claims
that it "creates problems." That's not a statement of theory, it's a
statement of alleged present fact. And there is no example of this
having actually happened, and, as I showed, what are asserted as
problems aren't.
He thinks of "preferred systems." The real preferred system, for me?
That people make decisions collectively, using the best methods
available to them, and based on carefully considered examination of
evidence and argument. If it were not for the problems of scale, I'd
prefer simple, direct democracy, with, then, the whole panalopy of
tools available under deliberative rules such as Robert's Rules of
Order. We should consider how to emulate this on a larger scale, so
that results would be similar.
Richie has insisted on his claim that RRO "recommends" IRV. In fact,
they recommend repeated balloting, and only acquiesce, obviously
reluctantly, when repeated balloting to find a majority is not
possible. Almost all the communities where some form of IRV has been
implemented, repeated balloting was, indeed, possible, it was being
used and had been used for years. So, to a community like that, to say
that RRO recommends IRV for them is so misleading that it's tantamount
to a lie, particularly when repeated, as it has been, after the
distinctions have been pointed out.
Given that the method described by Robert's Rules actually does
continue to find a majority, or the election is nevertheless, no
matter how inconvenient, repeated, FairVote is pulling bait and
switch, to convince people that a product is good, based on a
recommendation of it by someone reputable, offering your own product
as a cheap substitute, relying on recommendations of the better
product to help convince people to buy yours. Typical sales behavior,
and only marginally illegal in some places.
When RRO says that the method of preferential voting they describe is
"better than election by plurality," they are referring to the
contingency that a true majority found through vote transfers being
than electing without a majority. They are not saying that an IRV
plurality is better than a Plurality plurality, they are both
pluralities and both are prohibited elections under Robert's Rules,
unless an organization has decided that a result *must* be found with
one ballot and has explicitly allowed election by plurality. This is
not a problem in many organizations where the choice does not involve
a major change in direction.
And then Robert's Rules points out two major flaws of the STV method
that Richie calls "IRV." It has these flaws even if a majority is
found.
It's quite possible to read the preferential voting section in
Robert's Rules of Order, superficially, and walk away thinking that
RRO recommends "IRV." But that's dependent upon a quick judgment that
what RRO recommends is the same as the IRV that is being considered,
because of the similarity of counting, all except for how the final
conclusion is determined, a detail easily overlooked. So Richie is
right, lots of people have concluded that what he's saying is true.
But I've never seen a non-salesman stick with that conclusion after
the details of the text are pointed out.
Does this matter? In the end, any committee considering voting systems
should not depend on "recommendations," alone, but should examine the
basis of the recommendations. RRO is making very general
recommendations that apply to deliberative bodies and membership
associations. Your situation may be quite different. Richie advocates
IRV practically without reference to the individual communities
involved.
His home town has adopted IRV, where it is a fish bicycle. Small town.
Most elections are uncontested, a few with two candidates, and a rare
one with three, where IRV makes no difference at all. But it was
another feather in his cap! One more "implementation" that he can
claim, showing "momentum." Don't be snowed. If you are considering
IRV, or any election reform:
1. Consider how it is likely to work in *your* community, and if you
can find implementations in similar communities, look at the actual
voting patterns and results. Don't compare partisan elections with
nonpartisan ones, they behave quite differently. Don't compare
elections with two or three candidates with elections with over
twenty, as in San Francisco. Or at least understand that results for
one kind may differ from results of the other kind.
2. What is your existing system? What are its strengths and flaws? How
does the system influence voting behavior?
3. Beware of knee-jerk assumptions that existing behavior is Bad.
There is usually a reason for it. Salient example: low turnout in
elections. Voters don't vote when they don't believe the results will
make a difference in their lives. If they don't care if the result of
the election is A or B, the normal frontrunners, they have no
incentive to vote and all the efforts in the world to get them to turn
out to vote will only shift the election toward results based on pure
name recognition and very superficial media impression. In other
words, results shift away from being based on the opinions of people
who actually know and care, toward opinion from heavy media influence.
I'm just pointing out the possibility that the idea that "more is
better" with turnout may not be completely true! The goal of good
voting systems should be increased voter satisfaction with the
results, so that there is increased voter consent to being governed
and communities move toward cooperation rather than conflict. Actual
turnout may or may not foster this, it depends on conditions, and
probably working on campaign finance reform or ways to find general
consensus in nomination processes will do much more to good effect
than fixing relatively small flaws in voting systems.
4. Consider the alternatives. FairVote is offering you an expensive
car, that has been well-marketed and there are some existing sales. Is
it possible that there is a better car that costs less? Is it possible
that the car you already have, which perhaps has some problems, could
be fixed for far less than the cost of a new car? Richie claims that
allowing votes to be simultaneously considered has "problems."
However, if we look at the implications of this, every voting system
that would function to find the "compromise winner" which RRO clearly
thinks desirable, will consider votes simultaneously. He is
essentially saying -- and you will find him arguing elsewhere -- that
finding a compromise winner is undesirable, that we should, instead,
choose the candidate with the strongest 'core support," presumably
because this candidate will have an army of supporters to make his or
her government stronger. He's even invented an election criterion, the
Core Support Criterion, that is completely preposterous, it's simply
designed to make IRV look good, because, the way he states it, it
sounds desirable. It isn't, and any election method which satisfies it
is designed to fail under some circumstances, and badly.
5. Remember the basic democratic principle of taking no collective
action without the explicit consent of a majority. In direct
democracies, this principle is required by the rules. We abandon it
only for reasons of convenience or practicality with large-scale
democracies, but it remains the goal, or should. Our larger
democracies fail to realize the core vision of democracy when we
completely leave this goal behind, and seek only victory for our
faction, hang the majority. We need more people to work for this
majority principle, to make sure that our large-scale processes do
approximate what we would decide if we could deliberate and vote
directly. Further, the more that the majority will seek consensus
beyond mere majority, when possible, our sense of "ownership" of the
government will increase, our sense that governmental decisions are
"our" decisions, participation will increase, and this will, overall,
benefit everyone, because good decision-making is not a zero-sum game.
I may lose on this issue, but gain on that one, and better
decision-making means that, overall, I generally and usually gain.