Wednesday, November 16, 2011

Fortin moves into second in endorsement race

Jean-François Fortin has moved into second in the Bloc Québécois endorsement race, but Daniel Paillé stays in front thanks to the support of four-term MP André Bellavance.

Meanwhile, the addition of union support into the endorsement rankings has inflated Peggy Nash's numbers, though she still ranks third.

As always, right-click and open the leadership charts below in a new window in order to magnify them.

With the addition of several former MPs and a swathe of Parti Québécois MNAs (primarily from eastern Quebec, provided to me by the Fortin campaign), Fortin now has 34.5 points in the endorsement ranking (36.7% of all available points) and has vaulted ahead of Maria Mourani, who has fallen to 20.7% of the total.

But Daniel Paillé is still leading, thanks to the support of Bellavance. Paillé now has 40 points, or 42.6% of the total. This indicates a first ballot win for Paillé is no sure bet.

For the New Democrats, there have been no new endorsements but what has changed is the addition of union support.

Unions make up an important part of the NDP's support base, and many union members are also members of the NDP. Though unions won't have a set portion of the ballot as they did in 2003, their influence will nevertheless be important.

Peggy Nash has the support of the presidents of two unions: the Canadian Auto Workers and CUPE Ontario. Brian Topp has the support of the United Steelworkers.

After looking into the 2009 Ontario NDP race, I've decided to award 0.06 points to every 1,000 members of a union that is endorsing a candidate. I'll get into why I did that a little further on in the post.

But thanks to this addition to the rankings, Peggy Nash is now in a far more impressive third place.

Brian Topp still leads with 181.5 endorsement points, or 42.8% of all available points. Thomas Mulcair is second with 104 points, or 24.5%, while Peggy Nash is third with 81.5 points, or 19.2% of the total.

There have been no other major changes. Robert Chisholm, Nathan Cullen, Paul Dewar, and Niki Ashton all have between 10 and 15 points, putting them at between 2.4% and 3.5% of the total.

Now why the need for a change? First, let's look at how the endorsement system would work for the three most recent provincial NDP races. For the purposes of these races, I have treated MLAs and MPPs the same way as I do MPs.

Let's start with British Columbia, which is the most problematic (especially for the BC Liberal leadership race).
I jumped ahead to the second ballot as Dana Larsen had no endorsements and so dropping him off the first ballot made no difference.

As you can see, Mike Farnworth was the favourite to win. He actually was during the campaign as well. He had more caucus and federal support than either Adrian Dix or John Horgan, which led to him being favoured in the endorsement rankings. Nevertheless, the final result on the third ballot was close.

The Manitoba leadership race, however, worked much better with this system.
Greg Selinger had, by far, more support within the Manitoba NDP caucus than did Ashton, and the system overstates his support. But it does see him winning easily on the first ballot - thanks in part to the support of CUPE Manitoba.

In Ontario, both Gilles Bisson and Andrea Horwath had some union support, while Peter Tabuns had the support of Ed Broadbent. In the end, it seems that union support tipped the scales in Horwath's favour, as well as her good support within the caucus.
Though the system overstated Tabuns's support in the first and second rounds, the system worked well enough. After the first ballot, Michael Prue threw his support behind Bisson, while after the second ballot Bisson threw his support behind Horwath. That was the clincher, but without including union support the system would have suggested that Tabuns would have won on the first ballot.

Going through the permutations found that 0.06 points per 1,000 union members yielded the best result. At 0.05 points per 1,000, Tabuns still would have led on the first ballot.

The chart on the left lays out the points system, and now includes union support. Obviously, this is really only an important factor for the New Democrats (though the FTQ could get involved in the Bloc race). I think it will help the rankings and they look, to me, pretty plausible. But the caveats described in the first post on the endorsement ranking system still apply - especially since the NDP and Bloc will be using a OMOV system, meaning the results in December and March could look very different.


  1. Needless to say the BC Liberal contest is a major exception to these rules. Christy Clark won despite having the backing of only 1 MLA out of 49!

  2. Yep, but polls showed her to be much more competitive. It's the same situation with Alison Redford, so I'll be looking forward to some opinion polls on the NDP leadership race once things shake out a bit to see how the rankings work.

  3. As I recall there were a few polls of the general during the NDP leadership race in 2002/03 - they all had Layton third behind Blaikie and Nystrom - and all three way behind "DK/NA". It meant nothing because it was just a measure of name recognition and Layton simply had not been a national political figure for as long as the other two. Of course he won on the first ballot!

    There could be some interesting polls on the NDP leadership in February after several televised debates etc... but right now a poll would only tell us that Mulcair is well-known in Quebec and that the others are all unknown.

  4. hey there,

    There were a SLEW of NS provincial NDP MLAs and cabinet ministers at Robert Chisholm`s campaign launch. One only needs to view the footage of his speech to see the MLAs standing behind him and he thanked them and individually introduced them. Just wondering why they weren`t including in your endorsement system?

  5. Here is an article to support my claim about the Chisholm endorsements.

    "Dexter was joined by a number of his Halifax-area MLAs, including cabinet ministers Graham Steele, Dave Wilson, Maureen MacDonald, John MacDonnell, Marilyn More and Bill Estabrooks.

    They, along with backbenchers Becky Kent, Mat Whynott and Sid Prest stood behind Chisholm during the announcement."

  6. Thanks, I will include them.

  7. This is very impressive work, Eric! This basis of measuring leadership candidacy support offers an intriguing gauge of support, and a means of assessing support levels in the absence of much polling among party members and party supporters about their preferences.

    Any idea how your endorsement point model would apply to the federal Progressive Conservative race of 2003, or the Conservative Party leadership race of 2004? Also, have you noticed if your model is better at predicting the outcome of one type of leadership contest election system over another (for instance, a delegated convention VS. straight one member, one vote VS. weighted constituency voting)?

  8. junkiepolitico,

    I don't have information on endorsements for the 2003 PC and 2004 CPC races, so I have no idea.

    The system works better when delegates choose the leader, but it also works with OMOV or a mixed OMOV/delegate method. It would have been right for the 2003 NDP race, the 2009 Ontario PC race, the 2009 Ontario NDP race, and the 2008 New Brunswick PC race, for example, which all used OMOV to some extent.


COMMENT MODERATION POLICY - Please be respectful when commenting. If choosing to remain anonymous, please sign your comment with some sort of pseudonym to avoid confusion. Please do not use any derogatory terms for fellow commenters, parties, or politicians. Inflammatory and overly partisan comments will not be posted. PLEASE KEEP DISCUSSION ON TOPIC.

Note: Only a member of this blog may post a comment.