Sunday, November 25, 2012

The trouble with dynamic strategy canvases: Calgary Centre by-election

Calgary Centre is holding a by-election on November 26, 2012. A group of progressive citizens under the banner of 1CalgaryCentre (website and Facebook) has attempted what progressive political parties have been unable to do: create a process to select a consensus progressive candidate. It's a grand experiment; it is idealistic and optimistic. It suggests doing something is better than doing nothing.

That process returned valuable insights but that's not what this post is about really. (Yes, Chris Turner running for the Green party won - his campaign was way more effective at turning out the vote online).

Some people have been critical of the lack of "representativeness" of the 1CalgaryCentre process. Well, it wasn't intended to be representative - or predict an election outcome based on a 'snapshot-in-time' poll. Rather it sought to crowdsource a consensus candidate among engaged progressive voters, i.e. people likely to actually cast a vote. That is a fundamentally different objective than the typical election polls reported on in the media as if they could predict the outcome on election day (Yes, horse races can be fun to watch. Yes, voters benefit from having access to non-partisan polling as it shows them a snapshot of what people in their riding are thinking).

1CalgaryCentre is another, a unique data point progressives in Calgary Centre can consider when they go vote.

A worthwhile point relating to random election poll is that, yes, they are based on the science of statistics and yet, they consistently and dramatically over-estimate voter turnout. So, the real question isn't random polls and whether they are accurate: they usually are within the errors they measure at that moment in time (sampling error is measured by "margin of error", all others usually ignored in reporting - such as people inaccurately reporting about actually going to vote; it's human nature: voting is seen as socially desirable, so when Canadians are ask, they are prone to say they will vote.)

Actual voter participation comes in at about 60% on average in recent Canadian general elections; it was 55.6% in Calgary Centre in the 2011 federal election. The random polls published over-estimate turnout significantly. Return-on-Insight reported that 12% said they won't vote (thanks for the honesty!) and 16% were undecided. Similarly the last Forum poll on November 17 surveyed 403 Calgary Centre residents and 374 indicated who they would vote for, resulting in only 8% indicating they aren't voting. Matching the actual turnout in Calgary Centre in 2011, the true number of "won't vote" should be closer to 45%.

In essence this means that  the strategy canvas on which this by-election is happening is highly dynamic: voter turnout alone determines the result. The reason for this is that the random phone surveys suggest that 2/3 of Calgary Centre voters support progressives. It's actually mathematically possible for progressives to come in 1st and 2nd.

So here's some math for you:
Scenario 1

2011 voters Nov 17 Forum poll breakout 2012 by-election votes differential to leader
49235 35% 17,232 CPC 
49235 30% 14,771 -2,462 Liberal 
49235 25% 12,309 -4,924 Green 
49235 10% 4,924 NDP 
100%
Scenario 2
2011 voters (minus 10%)
44312 35% 15,509  CPC
44312 30% 13,293 -2,216 Liberal 
44312 25% 11,078 -4,431 Green 
44312 10% 4,431  NDP
100%
Scenario 3 
2011 voters (minus 20%)
39388 35% 13,786  CPC
39388 30% 11,816 -1,969 Liberal 
39388 25% 9,847 -3,939 Green 
39388 10% 3,939 NDP 
100%


You see how the lower the turnout, the more important the progressive vote becomes, ie the victory of the poll-leading party becomes ever thinner in terms of number of votes difference.

In essence, the strategy canvas is dynamic and not set (this is not a zero-sum game), so here is another calculation:
Imagine Green party supporters identified by the Forum polls, turn out to vote at 75%, while the other parties deliver on the high end of a by-election turnout:

2012 Forum poll eligible voters 2012 votes
Chris Turner (Green Party) - 75% of his 25% support25% 88520 16597
CPC (based on votes from scenario 2) 15509
Liberal  (based on votes from scenario 2). 13293
NDP  (based on votes from scenario 2) 4431
New base 55363

That's right. This would boost the overall voter participation rate quite a bit. Note, this calculation uses two different base numbers. And why not? This is a dynamic field, where variables move based on actual behaviour. It's within the Turner team's grasp to win. Needless to say, it's also within the Liberal party's grasp to win. Deeper analysis by others suggests that the chances of the Liberal Party campaign team boosting up support is less likely. Still, if both Turner and Locke were to mobilize their vote at high enough rates, Calgary Centre could return the most unexpected: a 1st and 2nd place for progressive candidates.

Confidence is a key to the scenario. People do like to vote for winners. That means, Turner voters have to feel confident they can win and then they might vote for him. That's the whole secret to momentum and translating momentum into votes. 

Those voting for Chris Turner must turn out at high rates; if they do they can secure a seat in Ottawa. Amazingly, that doesn't just mean people under 40 who would make the difference, but people of all ages. Random polls are reporting strong support for CPC among 18-34 which is plausible enough.

If we accept that Conservative voters are split by the political leaning of their particular candidate toward the far-right leaving the so-called Red Tories out, so that they either will not vote or vote Green (James Harris was a Red Tory who became leader of the Green Party), then Canada has the makings of a truly historic moment today, driven by citizens who are both engaged politically and fed up with being discounted.

Pathways to House of Commons for progressives
Calgary Centre progressives have a beautiful process here (despite its many limitations; most notable no cooperation by progressive parties) that says Turner has the greatest momentum and looks to be able to turn out the vote most effectively. The math says: show up and vote and a progressive will win.

If you were part of the 45% who didn't vote in 2011 - that's 39,000 people - realize it's up to you. The relatively small but critical number of NDP supporters also have it in their hands to elect a progressive by voting for Turner or Locke. The Greens can do it by themselves if they turn out in unprecedented numbers for Chris Turner. These are the three main pathways on a fluid strategy canvas. As you can see, turnout, which truly is a measure of enthusiasm, is everything.

With this I await the returns from tomorrow's vote to find out whether Calgary Centre isn't just progressive in spirit, but filled with enough progressive voters to send a progressive to Ottawa to represent them.

And I applaud the 1CalgaryCentre team on persevering throughout this first-of-its-kind citizen experiment. It has made this by-election worth watching from afar.


(Full disclosure: as a tax-paying, long-time permanent resident in Canada I cannot and do not vote in any Canadian elections. I grew up in West-Germany, a democratic country with a proportional and representative voting system - one vote for a candidate in the riding, one vote for a party, which are combined to create German parliaments. Half the parliamentarians are elected directly, the other half via the party lists, which usually take geography into account. It's a purposefully designed multi-party system that more often than not results in coalition governments, with coalition contracts forged during coalition negotiations following elections.)