UK moves closer to electoral reform. What if we had Alternative Vote in Canada?

On Monday, the British House of Commons passed a Bill setting out a May 5th referendum on the Alternative Vote (AV) system by a vote of 328 to 269. There was some talk that renegade Tory backbenchers may vote against it in opposition to the date scheduled for the vote (to be held in conjunction with devolved elections in Scotland and Wales).

The Bill would also reduce the number of seats in the House from 650 to 600 and equalize all constituencies by population.  Currently, many rural constituencies have much larger populations than urban ones giving the Labour Party an advantage.

The Alternative Vote system is an electoral system similar to that used in Australia.  Voters rank all the candidates in their constituency and a candidate is not elected until he or she receives a majority of votes in the constituency.

In Canada, the push for electoral reform was set back with convincing referendum defeats for reform in British Columbia and Ontario.  Federally, there has been little talk of electoral reform most likely the result of minority government.  With the public’s distrust of coalitions, the prospect of continual minority government is not arousing much excitement.

But what if Canada adopted the Alternative Vote system being proposed in Britain and used in Australia (Tom Flanagan has advocated for its adoption in the past see Everitt and O’Neill 2002).  Harold Jansen looked at the historical use of AV in the Prairie provinces and found it had little effect on the proportionality of the results (Jansen, Canadian Journal of Political Science, 2004)  It is difficult to predict the outcome with much certainty but I thought I’d give it a try using data from the 2008 Canadian Election Study.

Methodology

Using the 2008 Canadian Election Study I determined the aggregate 2nd choice preference for voters in each province.  As a proponent for looking at the local aspect of Canadian politics, this goes against my principles.  But it’s the only way to assess the potential impact of the system.  It also assumes that voting behaviour will stay constant in a different electoral system which is certainly not the case.

I am only interested in constituencies where the winner’s vote share was less than 50%.  If a candidate receives 50% of the vote in AV, he or she is elected regardless of other preferences.

Results

Acknowledging the limitations of this analysis, under an AV model there would have been some changes in the election results that could have had a profound impact on how Parliament worked and the likelihood of a coalition.

The big losers under AV would be the Conservative Party (-20 seats) while the big winner would be the Liberals (+20).  Theoretically, this makes sense as the centre-left of the spectrum has a greater proportion of voters.  Even when I broke 2nd choices out by province, the trend was fairly similar across the country.  Also, since voting behaviour is so regionally based, and in Canada, most constituencies are two party contests, AV should only have an effect in a small number of constituencies.  In my little simulation, only 26 constituencies would have had a different result on under the voting system.  The Conservatives were so dominant in many of their held ridings that preferences wouldn’t have come into play.  Only in Atlantic Canada and Ontario were some splits meaningful in changing the outcome.

The Canadian Election Study found that New Democrats were more likely to select the Liberals as their 2nd choice, while Liberals were more likely to select the NDP.  Conservative voters split 50/50 between the NDP and Liberals, while Green voters split 50/50 between the Liberals and NDP.

The national breakdown was:

2008 (actual) 2008 (w/ AV) Change
CP 143 123 -20
LP 77 97 20
NDP 37 42 5
BQ 49 43 -6
GP 0 1 1
IND 2 2 0
308 308

Provincial breakdown:

Atlantic
2008 (actual) 2008 (w/ AV) Change
CP 10 4 -6
LP 17 21 4
NDP 4 5 1
BQ 0 0 0
GP 0 1 1
IND 1 1 0
32 32
Quebec
2008 (actual) 2008 (w/ AV) Change
CP 10 9 -1
LP 14 20 6
NDP 1 2 1
BQ 49 43 -6
GP 0 0 0
IND 1 1 0
75 75
Ontario
2008 (actual) 2008 (w/ AV) Change
CP 51 42 -9
LP 38 46 8
NDP 17 18 1
BQ 0 0 0
GP 0 0 0
IND 0 0 0
106 106
Prairies
2008 (actual) 2008 (w/ AV) Change
CP 50 49 -1
LP 2 2 0
NDP 5 6 1
BQ 0 0 0
GP 0 0 0
IND 0 0 0
57 57
BC and North
2008 (actual) 2008 (w/ AV) Change
CP 23 20 -3
LP 6 8 2
NDP 10 11 1
BQ 0 0 0
GP 0 0 0
IND 0 0 0
39 39
David Coletto is CEO of Abacus Data and leads its Public Affairs research practice. He has a PhD from the University of Calgary and is an adjunct professor at Carleton University.

Contact David Coletto:

T: 613-232-2806 x. 248

E: david@abacusdata.ca

W: http://www.abacusdata.ca

Sean on Twitter David on LinkedIn David on Google Plus