The Historical record of effect of vote transfers on results
In elections where Alternative Voting was used, few seats were won by those who were not leading in the 1st Count. Therefore the results under AV were as disproportional as FPTP elections.
Alberta used AV to fill most of the seats, in prov. elections from 1926 to 1955.
AV Seat-winners different from FPTP winners
(Instances where the leading candidate in first count, the person who would have won if the election had used FPTP, did not win)
1926 2 (government lost one and gained one)
1930 (all winners same as under FPTP) (one seat later changed through recount)
1935 (all winners same as under FPTP)
1940 2 (government lost one and gained one)
1944 (all winners same as under FPTP)
1948 (all winners same as under FPTP)
1952 (all winners same as under FPTP)
1955 4 (government lost four)
So with total of about 560 seats filled through AV, only 8 seats were filled differently under AV than would have been the case under FPTP
Manitoba
Manitoba used AV to fill most of the seats in prov. elections from 1927 to 1953.
AV Seat-winners different from FPTP winners
(Instances where the leading candidate in first count, the person who would have won if the election had used FPTP, did not win)
1927 2 (Government party (Progressive party) gained two seats) (one other seat later changed through recount)
1932 1 (turn-over in St. Boniface - Cons. leading in 1st Count, Labour won in end)
1936 1 (turn-over in St. Boniface - Lib.-Prog. leading in 1st Count, Labour won in end)
1941 (all winners same as under FPTP)
1945 (all winners same as under FPTP)
1949 (all winners same as under FPTP)
1953 1 (Minnedosa turn-over - Lib-Prog leading in 1st Count, SC won in the end)
with maybe 300 seats filled through AV, only 5 were filled differently under AV than how they would have been under FPTP
British Columbia
1952 BC election -- Alternative Voting
Despite Liberals blaming AV for their defeat or blaming AV for SC victory, AV had actually little to do with the Liberal/Cons coalition dropping from 39 seats to 10 (10 being the combined total seat count for two parties).
There were only seven turn-overs - only 7 places where the leading candidate in the 1st Count did not win in the end. The leading candidate in the 1st Count would have won in every other district if the election had been held using FPTP.
Under AV, the candidate leading in the 1st Count won in all but seven of the districts.
1953 BC election -- Alternative Voting
Only in five districts did the candidate leading in 1st Count not win in the end.
Turn-overs (where candidate leading in 1st Count did not win in the end)
Lillouet Liberal won
Nanaimo Conservative won
Oak Bay Lib won
Prince Rupert Lib won
Vancouver Point Grey C Lib won
(Vancouver Point Grey C means the third contest in Vancouver Point Grey - in 1952 and 1953, in the multi-member districts, each seat was filled in separate contest -- the post system.)
========================
In 1952 and 1953 BC filled 96 seats through AV.
Only 12 were filled differently under AV than how they would have been under FPTP,
With so few "turn-overs," there is no surprise that AV produces just as dis-proportional representation as FPTP.
The turn-overs, in every case, were not caused by drive for proportional representation overall but only by pursuit of majority representation in the district.
The AV vote transfers had little impact on the dis-proportionality of the 1st Count, the vote tallies that produce winners under FPTP.
Generaly the winner was one of the major parties anyway, and that party in many cases was not under-represented so did not deserve the win proportionally.
====================================================
for more info on effect of transfers in historical STV elections in U.S. and Canada,
see Harris
Does the transfer of votes change the results? p. 376 (#50)
(vote transfers seldom affect results but they are useful back-up measures)
p. 379 (#53) election of those in winning position in first count
in Winnipeg elections 1920-1928, 81 were elected
and only six were not in winning position in 1st count
six were elected who were in fourth position in a district's 1st Count.
etc.
================================
Here's some stats to show how IRV is not good
winners under IRV are in almost all cases the same as would be under FPTP
(That means the IRV result is just as dis-proportional as FPTP.)
Provincially
AB 1924 to 1956
With total of about 500-560 seats filled through AV, only 8 seats were filled differently under AV than would have been the case under FPTP
MB 1927-1953
with maybe 300 seats filled through AV, only 5 were filled differently under AV than how they would have been under FPTP
BC In 1952 and 1953 BC filled 96 seats through AV.
Only 12 were filled differently under AV than how they would have been under FPTP,
With so few "turn-overs," there is no surprise that AV produces just as dis-proportional representation as FPTP.
City IRV
London ON city election 2018
In every ward and in the mayor's contest, the candidate who lead in the first count was elected every time. So the winner would have been the same if the election had been conducted using FPTP (assuming that voters would have voted the same).
Such a system is not any great improvement over FPTP.. In fact the same people are elected as under FPTP in most cases anyway.
Woud Liberals still win 160 seats under AV as they did under FPTP?
Likely they would.
Would they win more than that number?
It actually seems they would not, judging by these numbers.
in the 178 ridings where Liberals did not lead (338-160), the person who was leading (either Conservative, NDP, BQ or Green) likely would go on to win the seat, not the Liberal.
For one thing even under IRV, secondary preferences are not referred to unless no candidate took majority in the first count.
In half or so of the districts a candidate did take majority in first count so (if voters vote the same under IRV as they do under FPTP) in almost all ridings there will either be no second count (hence no possibility of different result than FPTP) or the result in the second count will be the same as the result in the first count anyway, judging by past IRV elections in Canada.
===========================================
Comments