Instant-Runoff Voting in Alberta 1924-1956 ------ winner-take-all majoritarian style on the western Prairies, as Harold Jansen saw it

I look at Harold John Jansen's article

"The Political Consequences of the Alternative Vote: Lessons from Western Canada"

Canadian Journal of Political Science / Revue canadienne de science politique

37:3 (September/septembre 2004) 647–669

In unknowing support of a Lib-NDP vote swap if Yukon adopts IRV,

Jansen says about the 1955 Alberta election --  "Over 80 per cent of the transfers from eliminated CCF candidates went to Liberals ~up from 38% in 1948 and 58% in 1952! 

and the Liberals reciprocated with two-thirds of their transfers."

so he gave support to the idea that Liberals and NDP might help the other in cases when a vote cannot help their own party, if history is any judge.

1926 

votes cast in Alberta in 1926 was 183,000 overall 

Edm and Clgary had 39,000

that leaves IRV districts vote count = 144,000

Jansen says 42.8 percent plumped, which means 62,000 plumped

1955

votes cast in Alberta in 1955 was 378,000 overall 

Edm and Calgary was 148,000

IRV districts votes = 230,000

29.8 percent plumped means 69,000 plumped.

well you can see those two figures of plumped votes are identical but the percentages are different, so, you see, looking just at percentage is misleading.

1948 Jansen gives 30 percent as plumped.

only so many districts had transfers in 1948.

How much of this was CCF votes?

only so many of those contests saw CCF candidate eliminated.

so we have just 9 districts where CCF votes transferred:

Acadia              430 of CCF's 607 votes were transferred - not plumped.

Bruce     600 of CCF-Conservative combined 1100 votes were transferred, not plumped.

Camrose 800 of CCF-Conservative combined 900 votes were transferred - not plumped.

Clover Bar       700 of CCF-Conservative combined 900 votes were transferred - not plumped.

Grouard            250 of CCF's 578 votes were transferred - not plumped.

Lac Ste. Anne  1050 of CCF's 1400 votes were transferred - not plumped.

Leduc too big field of candidates to judge CCF transfers from info in my source (A Report on AB elections 1905-1982).  Wikipedia "Leduc district" is no help.

Redwater             500 of CCF-LPP combined 820 votes were transferred, not plumped.

Vermilion  500 of CCF-Conservative combined 1000 votes transferred, not plumped.

so perhaps his low figure of 30 percent plumping is accurate 

I see now I should have looked at a different election where he put a high rate of plumbing/exhausted.

i'll take a stab at 1944 where he says 64 percent were plumped.

not many districts saw transfers:

Lethbridge      1300 exhausted out of 6000 valid votes

okotoks             406 exhausted out of 6600 votes

pincher creek 800 exhausted out of 4900 votes

St. Albert         400 exhausted out of 4200 votes

St. paul           600 exhausted out of 4100 votes

Vegreville       400 exhausted out of 3800 votes

Willingdon     600 exhausted out of 4000 votes

so I don't see where he got his 64 percent figure for plumped/exhausted, unless he is looking just at votes to be  transferred and saying how many of them were exhausted.

but that seems iffy.

(And he says that only places to be sure that a vote is pumped is in the second round of counting when a vote is transferred and it s found to be non-transferable.

Of course, there is no way that 65 percent of the votes cast were found to be non-transferable in the second round of counting, so I don't see the basis of his 65 percent. In fact by the end of all the counts, only 3000 votes (out of 37,000 valid votes) were found to non-transferable, so plumped votes looks like they can be no more than 8 percent at most.

(Admitted we can't know that the votes cast in the first round for the eventual winner and runner-up were not plumped. But it seems to me that is just guessing to assume they were plumped, if that is what he did.)

Not all votes exhausted were plumped. In cases where four or more candidates are in the running, some may have been secondarily marked for a candidate who was already eliminated -- admittedly it was not very common to have four or more in the running.

If we use vote found to be non-transferable as potentially plumped (at least for sure the voter did not rank all the candidates), we see these figures:

2nd Round

Exhausted . Total Exhausted . Total valid votes

Tally Percentage Tally Percentage

1926 6402 14.6 44,000 approx.

1930 0

1935 0

1940 0

1944 0

1948 0

1952 0

1955 ?

by election, Average of percent.

by numbers:

====================================================================

1926

Here are the total number of exhausted votes in all the districts where votes were transferred.

1926 Votes in use at the end Total Exhausted Votes cast

Athabasca 814 382 1196

Bow Valley 2095 450 2545

Cochrane 1686 299 1985

Edson 2358 428 2786

Innisfail 2390 175 2565

Leduc 4003 571 4574

Lethbridge 3675 866 4541

Macleod 1315 89 1404

Medicine Hat 3188 647 3835

Pincher Creek 1388 254 1642

Red Deer 3165 376 3541

St. Albert 2057 516 2573

Vegreville 3919 828 4747

Victoria 2719 521 3240

34,772 6402 41,174

6402

84% 15.5%

For some reason, Jansen says 42.8 percent of votes cast under IRV in Alberta in 1926 were plumped. But this figure is about three times as large as the total percentage of exhausted votes in districts that had multiple rounds of vote count, which he says are taken as indicators of plumped votes.

In fact, 84 percent of votes cast were in use at the end, so no way almost 43 percent were exhausted.

==============================

===============================================

STV figures

EDMONTON

If we use vote found to be non-transferable as potentially plumped (at least for sure the voter did not rank all the candidates), we see these figures:

2nd Round

Exhausted . Total Exhausted . Total valid votes

Tally Percentage Tally Percentage

1926 10 1659 9 18,154

1930 0 1643 8 21,189

1935 0 1642 4 37,267

1940 0 2613 6 43,743

1944 0 2747 7 37,834

1948 0 1894 4 46,150

1952 0 4890 9 52,039

1955 ? [20,009]* 26? 76,544

by election, Average of 9 percent.

by numbers: 37,000 332,000

total: 11 percent.

*1955 no full report on the 1955 vote counts seems available. 20,009 is how many votes were not used to elect the winners. But some of that was likely the votes held by the last un-elected candidate whose votes may not have been transferred.

===========================