[EM] IRV inconsistency

[Forest Simmons]

Here's an inconsistency of IRV that I wish somebody had told me about
before I submitted my article to the Green Voice.

It is possible for a candidate to "win" every precinct without winning the
election.

In other words, if the "winner" of each precinct is calculated by applying
the rules of IRV to the ballots from each individual precinct before
applying IRV to the entire collection of ballots, it can happen that
candidate B wins in every precinct while some other candidate wins the
election.

This weirdness cannot happen in Approval, for example.

Here's how to construct examples of this:

Start with examples of two precinct elections and then create larger
examples by putting together many two precinct elections:

Precinct 1
3 ABC
4 BAC
6 CBA

Precinct 2
8 ACB
5 BCA
4 CBA

Note that candidate B wins in both precincts, but is first eliminated when
the ballots are combined since the B faction has the smallest number of
first place votes overall.

Precinct 3
22 ACB
12 BCA
11 CBA

Precinct 4
10 ABC
18 BAC
27 CAB

etc.

These four precincts (or similar easily generated pairs) can be duplicated
endlessly to create very large examples where all of the precincts are
won by B, yet B is eliminated in the first round in the IRV count for the
election as a whole.

Note that (lone mark) plurality does not suffer from this inconsistency.

Why is IRV considered better than plurality when it fails this consistency
test and also fails monotonicity?

Is it only that it allows more ballot expressivity and more or less
eliminates spoilage by tiny parties?

In what other area does it clearly beat plurality?

Forest