Precisely What We Didn’t Want
On the morning of May 3, about 60% of Canadians woke up with precisely the federal election outcome they didn’t want: a Conservative majority government.
But, at the same time, the outcome was filled with surprises:
- an historic breakthrough of the NDP, going from 4th party in the House of Commons with only 37 seats to Official Opposition (2nd party) with 103 seats;
- the same NDP taking 59 of Québec’s 75 seats, which is more than the Bloc Québécois’ highwater mark of 54 back in 1993 and 2004;
- the decimation of the Bloc Québécois to only 4 seats in the Commons, thereby stripping it of official party status;
- the unexpected defeats in their respective riding of Bloc Québécois leader Gilles Duceppe to the NDP and Liberal leader Michael Ignatieff to the Conservatives;
- the entry of the Green Party through party leader Elizabeth May in her riding in British Columbia.
The vote counting was barely over that many started talking of a “merger of the left” (that is, between the NDP and the Liberals) in order stop the vote splitting between the two that allowed the Conservatives to win a simple plurality of votes in some ridings. On Facebook, someone created a page upon noticing that there were 14 ridings where the race was so tight that just a bit more than 6,000 votes made the differrence between a minority or majority Conservative government.
As intriguing and compelling as that theory is, it is, in my opinion as someone who has studied election results very closely, a bit off the mark. It is true that such a slight shift could have made a difference, but it doesn’t take into account the historical trend in those ridings, plus merely adding the votes of the candidates losing to the Conservatives assumes that those votes are flexible and interchangeable. This approach made sense in the 1990s and early 2000s when the right had splintered from the Progressive Conservatives to the Reform/Alliance, as the sum of the two did represent the vote on the centre right. But there was never such a “divorce” between the Liberals and the NDP; the NDP did not spring out of the womb of the Liberals as the Reform/Alliance had of the PC’s.
So, what I really wanted to know, based on the election results of 2011 and 2008, is whether or not there was bleeding of votes from one party to another opposing the Conservatives that led to actual vote splitting between those two parties and resulted in the Conservtives to come from behind and win enough ridings to tip them into majority territory. Thus I devised a formula that:
- eliminated the ridings the Conservatives already held going into the 2011 election;
- eliminated the ridings the Conservatives won in 2011 with at least a 50% + 1 vote majority;
- found the percentage of votes that “leaked” from the third- to second-place party from 2008 to 2011;
- based on the total number of votes in 2011 in each riding, calculated the actual number of votes that moved from the second party to the third (and may have even placed the third party in 2008 the second party in 2011), and finally,
- if the sum of the above number and the actual number of votes received by the second-place party was greater than the number of votes received by the Conservatives, vote-splitting was deemed to have occurred.
Interestingly, I also arrived at 14 ridings, but a much higher number of votes that had split, namely 46,496. While that may deflate the bubble of adherents to what I’m calling the Facebook theory, it still represents a mere 0.32% of the 14,723,980 valid ballots cast.
My full analysis can be found here. It concludes that, of those 46,496 votes, only 1,273 moved from the Liberals to the Greens and the remainder moved from the Liberals to the NDP, thereby unseating the Liberal incumbent. Eight of the 14 ridings are in Toronto, where the Conservatives made major gains, but so did the NDP — all at the expense of Liberal incumbents. So, had voters in those eight ridings stayed with the Liberals instead of riding the NDP orange wave, the likes of Ken Dryden and Martha Hall Findlay would still be sitting members of Parliament.
In short, the bleeding of 45,223 votes from the Liberals to the NDP in 13 of those ridings, including my hometown of Moncton, and 1,273 from the Liberals to the Greens in one of those ridings, gave us a 166-seat Conservative majority instead of a 152-seat Conservative minority.
And, of course, if we had a workable form of proportional representation (i.e, a MMP or “mixed-member proportional” system as I prefer and as advocated by the 2004 Law Commission looking into electoral reform in Canada), the Conservatives would be nowhere close to a majority with their 39.63% of the popular vote nationally.
Caution is advised when considering these tables, as they are extrapolating from the actual data from the 2011 election which was a first-past-the-post mode of voting. It is entirely possible that, if given 2 votes, voters might choose a candidate from one party as their local MP and a candidate from another party as their regional MP. That said, the tables below consider the popular vote, retain two-thirds of the seats as FPTP for local MPs, and redistribute the remaining third of the seats for the regional MPs according to a formula that takes into account local seats won to assign a proportion of regional seats so that the final result is a closer reflection of the popular vote.
|Federal election 2011: Scheme A|
|Scheme A divides the country by its provinces and territories, but per the recommendations of the 2004 Law Commission that studied electoral reform, it divides Québec and Ontario into two and three regions respectively. Unlike the commission’s recommendation, however, the three territories are grouped as one region, thereby keeping 308 seats in the Commons.|
|Federal election 2011: Scheme B|
|Scheme B divides the country into nine regions of more or less 35 ridings per region. This scheme retains the divisions for Québec and Ontario into two and three regions respectively, as recommended by the 2004 Law Commission that studied electoral reform, and results in keeping 308 seats in the Commons. Atlantic (Atl.) is Newfoundland and Labarador, Nova Scotia, New Brunswick, and Prince Edward Island; Prairies (P-NU) is Manitoba, Saskatchewan and Nunavut; AB-NT is Alberta and the Northwest Territories; BC-YK is British Columbia and the Yukon.|