As the Liberal leadership vote was determined by weighing each riding equally, the weighted endorsement rankings were the most comparable to the final results. Though the system is not particularly meant to be predictive (how the establishment feels about a leadership race will not always mirror that of the membership), the rankings were quite close to the final results.
The only major difference was with Martha Hall Findlay (all others were off by 2.7 points or less), and understandably so. As she was the only contestant to have a recorded endorsement in Alberta, she was awarded all of the province's points. That artificially boosted her in the weighted endorsement rankings.
Overall, though, the rankings performed surprisingly well. And, I'll admit, I did not expect them to. I assumed that 80% was not realistic in a leadership race and that Murray and Hall Findlay would be able to garner more support. That they didn't, and that Justin Trudeau did get just over 80% of the points (and 78% of the votes), is rather remarkable. If the party wanted to show unity, they've done it.
The following is an attempt to quantify the on-going Liberal leadership race by tracking endorsements. ThreeHundredEight has had some success in the past with this method, providing a good indication of establishment support within a party that can often replicate, or come close to, final voting results.
In the 2011 Bloc Québécois race, Daniel Paillé led the endorsement rankings throughout and eventually came out as the winner. In the 2012 New Democratic race, Thomas Mulcair and Brian Topp were shown to be the two frontrunners and, on the day of the convention, correctly estimated that Mulcair would beat Topp on the fourth ballot.
But as the system tracks endorsements, which generally come from the party establishment, the rankings can also tell us something about the results after the fact. In 2012, the system under-estimated Nathan Cullen's support, but that may have been more of an indication that his ideas were more popular within the membership than they were within the party establishment. As he was suggesting co-operation with the Liberals and Greens, that might have been expected.
The endorsement rankings use a simple system of assigning a value to each endorser. These values came about somewhat arbitrarily, but in addition to having been used to some success in 2011 and 2012 the values were calculated based on previous leadership races, as indicated here.
Generally speaking, party leaders, premiers, and veteran MPs are the most valuable endorsements to land. The chart to the left breaks it down by endorser.
But the 2013 Liberal leadership convention will not be a one-member-one-vote system, at least not exactly. Each riding is valued at 100 points in the voting system so that each riding is equally weighted (riding associations with a large number of members in Toronto, for example, will have the same weight at the convention as ridings in rural Alberta). Accordingly, I will be calculating both the share of endorsements points each leadership candidate has nationwide as well as weighted by province. The latter will probably be a better indicator of first ballot support than the former, but the tracking of endorsements without weighting by province is probably a better indication of actual support within the party establishment.
Why use this system at all? It is one of the few ways to quantify a leadership race in Canada. This isn't the United States, where registered voters can cast a ballot in the primaries, and where polls are generally reliable. There are no primaries here, and polls can only ask all Canadians or supporters of the Liberal Party who they prefer. Neither of these are necessarily representative of how the membership feels (though the membership may be influenced by general perceptions of the race).
Endorsements can act as a proxy poll of party members, though. Firstly, endorsers are party members. Secondly, their endorsements can influence other members. But most importantly, I believe that endorsements are a reflection of what the party thinks. People give their endorsement for many reasons, but two are probably the most important: one, they believe the candidate is the best person for the job, and two, they believe the candidate has what it takes to win. If an endorser thinks the candidate is the best person for the job, the odds are that other members will agree. If an endorser thinks the candidate will win, the odds are that other members think so, too. And these people within the party establishment are likely to be keeping closer tabs on the dynamics of the race and have better information than we outside observers. If they think a candidate will win, they may know something we don't know.
The endorsement tracker, then, almost acts as a reflection of the group-think. That can sometimes be self-reinforcing, but generally speaking the consensus opinion will be a good one. There are always exceptions (Christy Clark won with almost no establishment support), but the endorsement rankings can be a good way to keep tabs on the race. At the very least, it gives a strong indication of what the party establishment thinks. If the membership disagrees, then that in and of itself will be a result.
|As of April 12, 2013|
|Click to magnify|
Candidates are listed alphabetically, while endorsers are broken down by category and then listed alphabetically.
As mentioned, the leadership convention will not be giving each vote equal weight. Instead, each riding is assigned 100 points and each candidate will receive a share of those points based on their support in each riding.
The following chart shows the point share of each candidate by province. (The chart is styled after the graphics used by CNN during presidential primaries, for kicks.)
|Click to magnify|
|As of April 12, 2013|