top of page
Tom Monto

London Ontario 2018 AV election - good and bad

Updated: Feb 17, 2021

London Ontario held its 2018 municipal election using referential ballot. This is the first time preferential voting was used in a Canadian government election since Calgary's last STV city election in 1971.


The Alternative Voting system was used for the election of London's mayor and the election of its 14 councillors in 14 wards.


No proportionality was created by AV, but instead it ensured that to be elected a candidate had to have the support of a majority of voters in each ward (or in the city as a whole for the election of mayor).


However the city put a limit of two back-up preferences on each ballot.


This was only an issue if there were more than three candidates.


Mayoralty election

for the election of mayor an unusually large number of candidates ran - 14. By the time Holden was elected on the 13th Count, all but two candidates had been eliminated. Only 13,000 votes separated him from the other surviving candidate. By that point 21,000 votes had been exhausted. Perhaps as many as a fourth of them had been marked with only one preference, but with 11 vote transfers preceding the final count, many had exhausted even three choices if the voter had marked even that many.


Put another way, Eric Holder won in the end. He was the leading candidate in the first count, taking 34 percent of the first-preference vote, and kept a lead (between 11.5k and 13.3k votes) over his nearest rival.

However, there were a significant number of exhausted ballots. (More than a fifth of the votes in each transfer were unable to move due to lack of marked back-up preferences.) This resulted in Holder winning with about 60 percent of the vote still in play at the end, but with no more than 46 percent of the votes cast in the first count.


At 13th Count, 75,434 votes were still in play (21,000 fewer than in the First Count).

Holder had 44,373 votes, 59 percent of it.

His sole remaining opponent had 31,061 votes, 41 percent of it.


The large number of exhausted votes at the end, at least partially created by the artificial limit on marked preferences, were about 22 percent of the total vote. This is a much larger portion than in the 1948 Edmonton STV election when 3 percent were exhausted by the time the city's final provincial seats were filled. There was no limit on the number of preferences voters could mark in Alberta's provincial STV system.


It is even possible that Mayor Holden's main opponent might have been elected instead if full preferences had been allowed.


Ward Elections

AV in ward elections made no change from the FPTP-like first count results. Except it proved that the successful candidate was in fact worthy of the post by majority support.


The leader in the first count was elected in the end in each ward.


in three wards only two ran so no vote transfer was possible.


In five others, a candidate received a majority of votes on the first count so no vote transfer was needed.


ln the other six, successive elimination of the lowest-ranking candidate and vote transfers were used to establish the one with majority support.


In five, the field had to thin down to only two before a majority could be accumulated by any one person.


Ward Nine stands out as a case where three candidates were still in the running when a majority was accumulated by one candidate. The opponents support even amassed together was not enough to overtake the leader. Vote-splitting and strategic voting (AKA deceptive voting) played no part in this election.


Ward 8 proves the worth of AV and the problem of having a three-preference limit. Nine candidates contested the seat. The leader in the first count. Lehman, had no more than 30 percent of the votes. with another just behind him. And in each subsequent transfer he did not receive the most votes. He eventually won with just 150 votes more than his opponent. 1900 votes by this time were exhausted.


The arbitrary three-preference limit thus it seems secured his victory. A vote that shifted from candidate to candidate to candidate was exhausted when required to be transferred again. How many of these might have gone to Matt Reid is unknown. It seems likely that at least 150!


STV with no limit on preferences allows full illumination of the voters mindset. The limit of three preferences prevents this, as well as negatively affecting representation.


If ballot marking is thought too complicated -- although it was not a problem in Alberta STV/AV elections -- there is the expedient of having candidates preset to whom their votes will go if they are eliminated. This was described in Sandford Fleming's 1892 book on electoral reform (see other blog).


Thank you for reading.

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

5 views

Recent Posts

See All

Comments


bottom of page