Last weekend the Conservative Party of Canada held a leadership election and selected Andrew Scheer as their next leader. This decision has important political and policy ramifications for our country but those will not be discussed today. Instead, I am going to look at the electoral system used since it is particularly fascinating. The system used for this leadership race involved a ranked ballot but with each of the 338 federal ridings carrying equal weight. Inside each riding, it was one member/one vote but each riding’s results were then normalized out of 100 points. The actual raw vote numbers have not been released, although they would be informative to those trying to determine if this is a better system than one member/one vote ranked ballots for leadership elections. Meanwhile, the analysis of the election itself presented on Canadian television and written media has been weak since without the raw votes the journalists and pundits don’t really know what happened. The results as they currently exist are available here
While looking at the leadership results, I realized that each riding’s raw vote totals can be determined by figuring out the only numbers that make those percentages possible on each round. I tried this for a few ridings and determined the raw votes for Newfoundland and Labrador. Those are below and followed by some comments on what they suggest for the national analysis. The first table is raw votes for the entire province by round.
Newfoundland and Labrador Leadership Votes
One of the first observations is that the raw votes, don’t correspond very well to the provincial riding points. For example, Andrew Scheer received twice the riding points on the first ballot as Michael Chong despite having fewer votes. This is due to large differences in the Conservative Party membership between the ridings. Although, Newfoundland and Labrador is a small province its ridings are quite different from each other. This is particularly evident in comparing the well-educated, cosmopolitan, wealthier urban residents of St. John’s East to the dispersed Innu of Labrador or the rural ridings outside the Avalon Peninsula. Below, I am noting the total votes per riding from the first round of voting.
Votes Per Riding, Round 1
|Coast of Bays-Central-Notre Dame||69|
|Long Range Mountains||55|
|St. John’s East||207|
|St. John’s South-Mount Pearl||130|
The most obvious observation is that raw votes do not correlate well with provincial riding points even in provinces whose federal voting patterns are relatively similar. Urban ridings have more Conservative members and cast a disproportionate portion of the votes in this election. The result is that candidate’s whose primary appeal is to urban conservatives underperformed their raw vote. In particular, this hurt Maxime Bernier and Michael Chong in Newfoundland. These raw votes confirm the movement of social conservative voters towards Andrew Scheer in later rounds while also indicating considerable movement of voters from the more secular candidates (Lisa Raitt, Kellie Leitch and Michael Chong) towards Maxime Bernier.
 I have slightly less confidence in the 13th round, since there are a few possibilities for how Avalon’s 60/40 split actually arrived. Thus, the 13th round could be off by 3 or 6 votes for Bernier and 2 or 4 for Scheer with both errors in the same direction. For the other rounds and ridings, these are the only numbers which are possible.