[EM] Does IRV elect "majority winners?"

Abd ul-Rahman Lomax abd at lomaxdesign.com
Sun Jan 4 11:07:31 PST 2009

At 02:29 PM 1/3/2009, Terry Bouricius wrote:
>I am extremely versed in state legislative parliamentary procedure, as I
>served ten years as a state legislator. Nearly every state uses Mason's
>Manual instead of RRONR.

That's right. RRONR was based on the practice of the U.S. House of 
Representatives. There are plenty of differences in detail. Every 
deliberative body, as a general rule, may set its own rules governing 
its deliberations. It's actually a loophole one could drive a truck 
through, but, it seems, it is not *much* abused.

(The "nuclear option" in the Senate involved the presiding officer -- 
possibly Dick Cheney -- ruling that a simple majority was adequate to 
cut off debate on consenting to a judicial appointment; since this 
was contrary to the rules, on the face, presumably there would be an 
appeal of the ruling to the body itself; if I'm correct, it takes a 
majority to overturn the ruling of the chair, but just a simple 
majority, not the 60% required for normal cloture. Needless to say, 
many Senators were concerned about precedent; cloture rules are a 
protection against ill-considered decisions. It's bad enough that, 
not too many years ago, cloture was reduced from 2/3 to 3/5.)

>   The basis for calculating majorities is specific
>to the situation, or constitutional provision. There are a few cases where
>the basis is the entire membership (such as over-riding a veto, etc.) but
>for most votes it is a majority of those present and voting -- assuming a
>quorum is present.

That's right. Deliberative bodies do not normally use any voting 
system other than simple majority rule, and, if I'm correct, absolute 
majorities are reserved for special situations. It would not be 
desirable for a veto override to be made by a minority of members, 
which can happen if only a quorum is necessary. This is probably an 
example of where the body is restrained by its charter, i.e., the 
state constitution, which, I'd imagine, requires that absolute supermajority.

>There are a few rare examples of national popular elections in which the
>basis is the entire eligible population, such as some referenda in Italy,
>where opponents often urge their adherents not to vote as a way of
>blocking the passage of some measure (since all of the abstainers who
>don't care to participate one way or the other are effectively counted as
>"no" votes).

This is desirable when a measure takes away essential rights. 
Otherwise it's a device for maintaining status quo, because of the 
difficulty of gaining an absolute majority. However, if such a 
measure is actually and seriously worthwhile, it shouldn't be that 
difficult to gain an absolute majority; it is effectively a 
supermajority requirement, declining in percentage as the turnout 
increases, to a minimum of an absolute majority.

>However the norm in governmental elections is to discount all abstainers
>from the basis, regardless of the manner of their abstention.

An "abstention," however, means someone who doesn't vote on a 
question. This is the only place in this post from Bouricius where he 
appears to be inserting a spin.

"Regardless of the manner of their abstention" is a loophole one 
could, again, drive a truck through, and, indeed, that's what 
Bouricius is attempting to do. There are two questions, here, and he 
will, if history is a guide, attempt to confuse them, because then 
the argument he will make for the first one, which is stronger, he 
can hope will rub off on the second. Debate tactic.

(1) If there is a preferential ballot, used by such a body, for an 
election or for any multiple-choice question, and a winner is 
determined by the rules, and members have abstained from making some 
pairwise comparisons, is this an "abstention" if they have voted in 
other pairwise races? It is clearly an abstention if they have voted 
in none. But, if they have voted for someone, one candidate at least, 
participating in the process, the usage of the term "abstention" 
becomes highly problematic. It is "like" an abstention, in some ways, 
but it is an "abstention" that generally represents equal-ranking 
bottom, i.e., a pure "No" on the election of all those not ranked; 
because the voter is utterly opposed -- one may assume by default -- 
to all those candidates, ranking them would imply approval of any one 
of them ranked, and could create a majority, and therefore the voter 
will be responsible for electing someone whom the voter might 
consider to be a monstrous choice. And this is how Robert's Rules 
will treat it, and Robert's Rules does consider the question. It is 
not an abstention, it's a vote, and it counts in the basis for 
majority. And I also think this is just plain common sense. Does 
Mason's Manual even consider the question?

(2) If the ballot does not allow full ranking, and a voter has used 
the available ranks, is the vote then an abstention against all other 
pairs? Given that Bouricius has already stated that he doesn't 
consider methods that don't allow full ranking to be ones which 
"guarantee a majority," the answer is clearly "No."

Now, a constitutional issue. Suppose we set up a method which 
examines all the exhausted ballots, and which determines if they have 
used the available ranks, according to some standard. (This has been 
tried in the U.S., and the rule was a little more complex than 
requiring full ranking, it allowed truncation, but where it was 
allowed depended on the number of candidates.) If a ballots does not 
adequately rank the candidates, this ballot is disregarded in 
determining the basis for a majority.

That is, this ballot can (more legitimately, but I still claim 
incorrectly) be considered as abstaining voluntarily from the last round.

Prediction: if this is tried, it would be found unconstitutional. 
Oklahoma Bucklin was found unconstitutional because it attempted to 
require additional ranking beyond first rank. (Contrary to what some 
propaganda about Bucklin would imply, the finding there had nothing 
to do with the method itself, and a dissent in the ruling claimed, 
quite cogently, that all the state court should do is to rule the 
particular provision null and void; that this was not done, given 
that the rule was not an essential feature of the method, but just a 
wild idea someone inserted (poison pill, anyone?), indicates that, 
big surprise, there are forces which don't want advanced voting 
systems implemented.

Because the rule would impair the participation of those who bullet 
vote, it would be, correctly, seen as an impairment of their 
privilege to make unconstrained choices.

It would simply be a somewhat milder form of requiring full ranking.

A majority requirement is a fundamental part of democratic process. 
It appears that it was common in early state constitutions for state 
elections; in, at least, some states, election by plurality was a 
*reform* that was only implemented in the 19th century. Why was it a 
reform? Well, because of the difficulty of holding elections, further 
process generally meant referring the decision to, say, the state 
House of Representatives or equivalent. It was considered better to 
keep the decision with the people. That's a highly questionable 
decision, and probably represents keeping the decision with the most 
easily manipulable, in some senses. Asset Voting handles this much, 
much better. It leaves the decision either with the people directly, 
or with those whom the people have freely chosen, without constraint, 
to make it on their behalf. If we had a truly representative 
Assembly, one that represents all the people with minimal compromise, 
then the *original provisions would work.*

More information about the Election-Methods mailing list