Top Two (non-instant) Runoff versus "Instant" runoff (executive summary)
In the "top two runoff" (T2R) system, one of the most common election systems in the world (2013)
there is a plurality-style
election whose top two finishers go on to a second round head-to-head
runoff election. (Unless somebody got over 50% in round 1, in which case most
places skip the second round.)
In the "instant runoff" system (IRV),
pushed by some as an improvement,
each "vote" is a rank-ordering of the candidates. We repeatedly eliminate
(both from the election and all ballots) the
candidate ranked top on the fewest ballots.
Eventually only one remains – the winner. In this way
the two (or more) rounds in a runoff scheme get simulated using only a single election.
Unfortunately the IRV system is more complicated for both voters and counters, which
often causes cost
increases (as well as sometimes-vast increases in
ballot-spoilage
rates)
despite the fact the second round is not needed –
and
it loses the benefit of the two-round system that the press will examine
the two contenders in a lot more detail between rounds, enabling voters to then
reconsider and make a more-informed choice.
We examined
the data from a large set of elections and found a very significant difference
between the 2-round systems and its instant "simulation":
the IRV winner is much more (about 3 times more)
often the same as the winner of the first round.
In other words, that "re-consideration" by the voters between rounds in
the two-round system, really
does make a big difference, and really is a big advantage that IRV loses.
Specifically, in IRV elections, the winner is the same as the first round winner 95% of
the time. With two round (genuine) runoff, the winner is the same only 84% of the time.
This is a very big difference, and hence we suspect it is well worth having that second round.
There is another
very significant difference between IRV and T2R (which may be related): the countries that use IRV
always (so far) rapidly develop massive and permanent 2-party domination in IRV seats.
Third parties find it virtually impossible to win, for example in Australia their
(IRV-elected) house recently contained zero third-party members 3 times in a row, despite every
seat being contested by more than 2 parties each time. (Do not confuse
their House with the Senate, which is not elected with IRV.) Austrialia's third
parties
all want to abolish IRV,
as do Australia's citizens.
However, the countries that use T2R, usually obtain a more healthy status
in which more than just 2 parties can and do win.
The USA pro-IRV propaganda group "FairVote" unfortunately has been able to
eliminate 2-round runoffs
in favor of instant ones in a number of places in the USA, most notably San Francisco.
This probably hurt USA democracy. It would be much better if they focused on replacing
one-round plurality elections by IRV (if they are going to push IRV, which is not
as good an idea as score voting) because that
far more clearly would represent
an improvement.