Leadership Projections

One interesting thing about leadership races is that you never know for sure how strong the respective candidates are doing. The media loves to go on about momentum and anoint frontrunners but the truth is that few journalists are really plugged in to the inner workings of the Liberal Party. What decides these races is not blog endorsements or the number of fluff pieces in national media but the number of Liberals who back each candidate.

While there were a lot of news stories about membership sales in July, there hasn’t been any sort of number crunching to try and gauge the relative support of different candidates. So, partly out of interest, and partly because I think it’s useful to have an idea of where the leadership camps sit, I decided to take a crack at predicting first ballot support for the different candidates. This took a lot of time and effort, so please be gentle.


Methodology

Right off the bat, I’ll say that there are obviously major flaws with the methodology I used because a lot of factors are very hard to objectively quantify and a lot of the numbers I’m using are best guesses. I’m not saying that what I’m projecting is what will happen at the delegate selection meetings but I would be confident to put these predictions up against any other predictions anyone wants to make (so if anyone wants to make a friendly wager, I’m game).

The goal is to predict the number of delegates each candidate will have vote for them on the first ballot. Right off the bat, we know that roughly 15% of the delegates are ex-officio. So, using the wikipedia endorsement page, I’ve determined the percentage of declared ex-officio delegates each candidate has in their corner and assigned these numbers to that 15%. While wikipedia isn’t perfect, after this CTV news story appeared, it’s obvious that almost all of the campaigns went to the site and added the names of their supporters to boost their numbers. So I’m fairly confident this will give us a good idea of ex-officio support.

The other 85% of the delegates are the tricky ones to determine. These will be selected proportionally, with Liberals on the membership lists as of July 4th eligible to vote. There will be two types of Liberals on these lists – new sign-ups and existing Liberals. New sign-ups are the individuals who were signed up by campaigns in recent months for this leadership race. Existing members are people who were already on the membership list.

Both officially and unofficially, many campaigns and Liberal provincial associations have released their numbers for new sign-ups. There was also significant media coverage about this in July which you can read here, here, and here. Using these numbers, blog buzz, and what I’ve heard, I’m able to predict another percentage of the delegates. For example, in Ontario there were 35000 old members and 45000 new sign ups. Reports are that Michael Ignatieff signed up 10000 new members in Ontario. Therefore, I can predict that this will give Ignatieff 10000/80000, of Ontario’s elected delegates. Once this is done for all the candidates, I will have assigned 45000/80000, or 56% of Ontario’s elected delegates to various campaigns.

So the question now becomes how to assign the other 44% of Ontario’s elected delegates, right?

After making phone calls and asking for feedback, I can get a general sense of where existing support is, but that comes across as fairly subjective on my part. So in my formula for estimating existing support, I’m only basing 1/5th of the total for each candidate on the perceived support the candidates have in that province. The other 4/5th I assigned based on a formula as follows (in each case, this is done as a percentage of the total. Bob Rae has raised 30% of the total money of all camps together, so he gets 30% * 20%, or 6% of this total).

20%: Money raised (since, like it or not, money is important in these things)
20%: Number of donations (this is a sign of grass roots support)
20%: MP endorsements (MPs can have a big influence on their ridings)
10%: Media mentions (I simply did a google news search for each candidate and used the number of stories that came up)
10%: Wikipedia endorsements (under the assumption that these are all somewhat important people with some influence)
10%: Blogger endorsements (since blogs give a good sense of what Liberals who follow the news closely think)
10%: Website (I just used Blue Blogging Soapbox’s most recent website rankings. If nothing else, it might give a sense of how good a communications team each campaign has)

The first three factors I weighed heavier because the numbers a lot less subjective (warning: these are the month old fundraising numbers - I'll update within the next few weeks with the new numbers). Media is important, but just counting news stories doesn’t give you an idea of how positive the news is, so I couldn’t bring myself to weight it for more than 10%.

Using this system, I was able to assign “existing member support”. Obviously this isn’t a perfect system, but I think it does give an idea of who has the momentum, air game, and support on the ground among long time Liberals.


Province by Province

If you’d like to see the numbers I used, click here. The following is merely a brief outline of the province by province support for those who think this post has already gone on way too long. I'll start out west, to alleviate alienation:

BC: Kennedy was first in new sales, followed by Rae, Dion, Volpe, and Ignatieff. The existing membership seems to be split between Kennedy, Rae, Dion, and Ignatieff. Public Eye Online has been very good at releasing specific numbers for this region so I’m very confident in the ones I used for the front running candidates.

AB: In a province where the membership list more than tripled, Kennedy was first in new sales. Ignatieff, Volpe, Rae, and Dion all had good sale totals and Scott Brison has a presence in some ridings. My sense is the existing grassroots membership is split between Kennedy, Dion, and Ignatieff, although both Brison and Rae have some very influential supporters.

Sask: Bob Rae led the way in new sales , doing very well in aboriginal communities. Kennedy was second and while Dion had nothing before, it seems about 200 of Orchard’s followers signed up, which will no doubt help him. Ignatieff didn’t sign up many new members but has good support among the existing Liberals. As is the case in Alberta, Martha Hall Findlay actually has a bit of support on the ground, thanks to her rural bus trips.

Manitoba: This is certainly Dryden’s strongest province. I’ve heard mixed reports about who, between him and Rae, sold more memberships but those two were certainly 1-2. Ignatieff was third in new forms sold, but probably has the most existing support (followed by the other two). Kennedy comes in fourth in his native province. Of note, Maurizio Bevilacqua performed well here, and in all the prairie provinces – I’ve moved his forms to the “existing members” list because I think it would be really rash to assume they’ll all follow him to Bob Rae. (especially given this)

Ontario: Ignatieff and Kennedy were tops for new sales, with Volpe in the mix. The next tier was Rae, Dion and Dryden. Another tier bellow them was Brison and Maurizio. It should be noted that the straw poll among Ontario Young Liberals at Spring Fling had Ignatieff and Kennedy neck and neck, with Dion a distant third and no one else on the radar. While this was just a poll among youth, a third of the delegates at the convention will be youth and there are over 200 delegates coming from campus clubs to Montreal.

Quebec: Volpe raised eyebrows with his Quebec recruitments numbers, although the figure of 4,400 new forms sold has been attacked as a dramatic inflation of his support due to their concentration in Montreal and questions about how loyal those Liberals are to Volpe. Ignatieff was a strong second, Dion was third, and Rae was fourth. Those three are all doing well among existing members too.

Atlantic Canada: I’ll be the first to admit that my Atlantic numbers are far from exact and I hope to be able to refine them over the next few weeks. The main problem here is that there were very few new sign-ups in comparison to the existing membership lists. Brison most likely won his home province (although some say Ignatieff did) and has decent support across the region. The top three in PEI are Ignatieff, Kennedy and Rae. New Brunswick is similar. The Toronto Star has Kennedy leading in Newfoundland, but other reports have Volpe, Ignatieff, Rae and Dryden all doing well there.

Territories: Unfortunately, I’ve ignored these three ridings due to a lack of information.


Results
Click here for the detailed results page and for how these numbers were calculated


The Ex-Officio support breaks down as follows:
Ignatieff 69
Rae 50
Kennedy 49
Dion 46
Dryden 34
Brison 29
Bennett 8
Volpe 7
MHF 4
Fry 2


The New Membership Sale Support Breaks Down as Follows (this has been weighted based on the relative delegates for each province and the ratio of new sales to existing members):

Iggy 20.5%
Kennedy 20.5%
Volpe 17.7%
Rae 15.9%
Dion 11.6%
Dryden 6.7%
Brison 4.9%
MHF 1.2%
Bennett 0.8%
Fry 0.2%


Click here to take a look at the data used to calculate relative strength of existing candidates among existing members.


Here’s how it breaks down for existing members once existing strength is considered (again, weighted by relative delegates for province and ratio of new sales to existing members):

Iggy 25.1%
Kennedy 17.1%
Dion 16.6%
Rae 15.4%
Dryden 9.4%
Brison 9.0%
Volpe 2.9%
Bennett 2.1%
MHF 1.9%
Fry 0.5%


Finally, the grand total of estimated first ballot support:

Michael Ignatieff 22.9%
Gerard Kennedy 18.4%
Bob Rae 15.8%
Stephane Dion 14.4%
Joe Volpe 8.8%
Ken Dryden 8.6%
Scott Brison 7.4%
Carolyn Bennett 1.7%
Martha Hall Findlay 1.6%
Hedy Fry 0.4%


Analysis

Obviously this isn’t a perfect system, but I don’t think it’s skewed to favour or punish a given candidate disproportionately. And, because of my Kennedy bias, I went out of my way to really estimate his numbers on the low end of what I’d heard whenever possible.

I’ll be updating this every couple weeks from now until the end of September, so feel free to e-mail me with feedback on the estimates I’ve made. The percentages will also shift as more endorsements become public and more up to date financial numbers come forward. I should also stress that I’ve been working on this over the past week, so some of my totals from the wikipedia page and google news search might be a few days out of date (as are the fundraising numbers).

As for the results themselves, they confirm what I’d suspected. Ignatieff is still the frontrunner and will be in first place on ballot one, but it won't be an insurmountale lead. If he is only around 25% on the first ballot, that means we’re heading for a long night of voting on December 2nd. Despite a lack of positive media focus lately, Gerard Kennedy should be in second place after ballot 1. I’m personally not too worried about the media’s lack of attention because that means he’ll be perceived to have a lot of momentum after the delegate selection meetings finish. But I do think it’s ridiculous that he’s at times being ignored when the front runners are listed.

As one might expect, the only other candidates over 10% are Rae (in third) and Dion (in fourth). Despite lackluster new sale numbers in Ontario for these two, they both did very well on the formula I created to estimate existing support and I think I was fairly generous with Dion’s numbers, due to his perceived momentum. Rae might also move a bit higher if he can corral some of Bevilacqua’s supporters his way.

The next tier of candidates includes Volpe, Dryden, and Brison. I’ll admit to be a bit surprised by this since I’d expected Volpe to do better, given his high sale numbers. Perhaps I was a bit harsh in my assessment of his existing support, but if the system I used was weighted against him, that will merely make up for the loss of Jimmy K. Because a lot of the new forms were Karygiannis' and not Volpe’s, it’s going to be tough for Joe to crack the top 4.

Unfortunately, the three female candidates seem to have little support. These numbers might be off by a fair amount because I really have no sale information on the low tier candidates. Carolyn Bennett could have been anywhere from 0 to 2000 forms in Ontario for all I know. I also strongly suspect that MHF is doing better than these rankings indicate. She just seems to have a certain je ne sais quoi which shows up in support, but not in these rankings. Also, I get the sense that a few ex-officios might give her their first ballot vote to make a statement before moving on to their preferred candidate.

In summation, this is far from an exact science, but I'm fairly confident that it provides a very good indication of what we should expect on the first ballot in Montreal. And, judging from the numbers I've projected, there are, without a doubt, four candidates who could realistically win this thing.