I went into some detail of Edmonton's experience with the Single Transferable Voting system of proportional representation in another blog -"Change Edmonton elections today! STV and at-large elections now!"
But here I want to explain the reason why Edmonton's STV experience was so short.
The guarantee of two southside councillors muddied the 1926 election.
As we see here in my summary of the 1923 vote count:
In the 1923 election 14 candidates ran for six empty council seats, all to represent one city-wide district.
When the votes were examined, one Labour and one business candidate were elected on first preferences by exceeding the minimum required to win a seat.
Transfers of their 2500 surplus votes then started the creation of formed consensuses. This continued through the elimination of seven lowest-ranking candidates, and one more Labour and one more business candidate were elected by the end of the tenth count.
At the time a guarantee was in effect that at least two candidates resident in the southside would be elected.
At the end of the tenth count, all but two seats had been filled, with none so far going to southsiders. So the two southsiders with the largest totals were declared elected to fill the two remaining seats.
However when these seats were allocated, the remaining northside candidates were simply disregarded - their votes were thrown away. Many charged that as they had been eliminated, the back-up preferences marked on the ballots should have been consulted and the votes transferred to the first-ranking southside candidate on each ballot if so marked.
There was also an on-going campaign against the at-large system (perhaps with dis-empowered powers-that-be behind it trying to undercut STV). But wards were not formally put forward as an alternative to STV. And in fact it is possible to combine STV and wards by having multi-member wards and voters only casting single votes. But it seems distrust of the at-large system was used to undermine STV.
The practice under STV that voters cast only one vote accompanied by back-up preferences was always a hard one to sell when compared to Block Voting where voters could cast as many votes as the number of open seats. Some said they felt they were losing voting power by only having one vote, but others pointed out that it was better to have one vote that was effective than six that were diluted among everyone else's six votes and then often wasted.
There were many spoiled ballots in Edmonton STV elections, but that number seems to include many ballots where a section was left blank, such as a voter voting for mayor but not for city council or for mayor and city council but not for a school board. They should still have been used for the sections filled out properly but may not have been.
As well, the staggered elections reduced the effectiveness of proportionality. Only half the seats were up for election each year so in 1923 there were only six open seats on city council, and only four in each of the public and separate school boards.
The guaranteed southside representation inhibited the election of labour or other minority representatives within this relatively small crop of representatives.
Thus for city council, a voting block of 14 percent was required to elect a councillor in the STV elections (excepting if additional seat(s) were empty due to resignations or death). With two seats going to southside candidates, it meant proportionally perhaps 20 percent was required to elect a candidate. That is, a group that only had 19 percent of the vote could not get seat, at least theoretically. The portion would vary depending on whether a southside candidate was supported by many votes from northside voters - votes could go anywhere, the rule was just that two southside candidates were guaranteed seats.
Quota, the number of votes required to win a seat, was calculated by dividing the total votes by the number of open seats plus one. Thus four seats could be filled by four groups, each having a little more than 20 percent of the vote, and slightly less than 20 percent of the vote could be wasted. This shows up badly compared to the provincial use of STV in Edmonton where five seats were filled by voting groups as small as 16 percent of the vote, or the provincial use of STV in Winnipeg (prior to 1949) where 9 percent of the vote was enough to elect one of the 10 city MLAs. There a Communist and two other labour parties were represented in the Legislature, as well as the larger Conservative and Liberal groupings.
So under Edmonton's STV, little actual minority representation was created despite the extra work involved in the vote transfers. It should be noted though that STV did not create more work generally. Block Voting meant counting up to six votes for each voter, while STV meant counting only one first-choice vote from each voter then many less for each subsequent vote transfer, perhaps as few as only a few hundreds in each count. So STV did not entail much if any additional counting. (The difference though was that STV counting was in different rounds, all but the first conducted at the central election office, while Block Voting counting is done simultaneously at each of the different polling places.)
In 1922 Block Voting 53,000 votes were cast (by 11,000 voters voting).
in 1923 STV election 13,000 votes were cast.
The votes were gathered at the central office.
Then were 9 transfers. (The first two moved 2500 votes, so altogether probably not more than 7,000 votes were transferred. This was less counting in total than under Block Voting.)
Altogether these factors resulted in pressure to hold a plebiscite in 1927 on returning to at-large Block Voting. A majority of voters went for change, and Edmonton's municipal STV passed into history.
it was replaced by Block Voting, and 78,000 votes were cast in 1928.
If transfers were the issue, STV could have been replaced by Single Non-transferable Voting (SNTV). Still electing multiple councillors at-large and still each voter wields only one vote. The difference being that the vote is non-transferable - votes are wasted if not placed with winning candidates. This system would have provided mixed representation, roughly proportional, even without transfers and despite the waste.
However voters casting multiple votes in at-large elections is taken as the norm so much that the term Block Voting is seldom seen. And Edmonton by instinct returned to Block Voting after STV's snubbing in the referendum.
See my other April 2020 blog on "Block Voting" for info on this oft-overlooked third system of district voting.
If Edmonton had adopted SNTV, there would have been guaranteed no way for any one group to take all the seats available in a city council election (unless it had support of 70-90 percent of the voters).
As it was, the voting blocks (Labour and business) were not solid, and it seems many gave votes to both Labour and business candidates. This could have been due to fact almost all those elected were former councillors with name recognition and several elected were sitting councillors going for re-election. Perhaps that trumped party lines.
Four Labour and two business candidates were elected in 1928. This was better than Labour had managed under STV.
If Labour is taken as a minority party, STV had not been actually that kind to it, despite the claim that the STV system nurtures minority representation:
in 1923 two Labour councillors were elected.
in 1924 two Labour councillors were elected.
in 1925 three Labour councillors were elected.
in 1926 two Labour councillors were elected.
in 1927 two Labour councillors were elected.
then in 1928 with the return of Block Voting, Labour elected four councillors!
Of course changing economics or other factors may have changed voting behaviour apart from the electoral system in use.
See my blog "Workers voting for workers" April 2020 for some remarks on this.
Thanks for reading.
===========================
Comments