Wednesday, January 5, 2011

2008 Alternate Scenarios #1: D'Hondt, Part One

Prince Edward Island farm rainbow 2009Image via WikipediaSo I've decided to look back at the 2008 election and crunch some numbers to consider how the votes cast in 2008 might have brought a different parliament had Canada had a different voting system than it currently does. I'm starting with the D'Hondt Method, a proportional representation system that delivers MPs in a ratio approximating the ratio of votes they recieved. It's a 'party list' system in principle, meaning you vote for a party and not an individual, and the MPs are chosen from lists. It might not necessarily have to be that way, but that's how it's applied generally.

Now, 2008 gave is the following result:
CPC 143, Lib 77, BQ 49, NDP 37, Ind 2.

A simple nation-wide proportional representation calculation using the D'Hondt method would have returned this:
CPC 118, Lib 82, NDP 56, BQ 31, Green 21.
Which is a huge decrease for the Conservatives and the BQ but a huge increase for the NDP and Greens. I'm surprised to find the Liberals actually gaining seats in a prop-rep simulation.

I can't really see a system like this flying with the Canadian people, mind you. It's too anonymous, and it sees Canada as one huge mass as opposed to a federation of provinces. It might go over better to apply the D'Hondt method to the results of each province (retaining the current number of seats per province). If we do that, we get the following:
Newfoundland and Labrador: Lib 4, NDP 2, CPC 1.
Prince Edward Island: Lib 2, CPC 2.
Nova Scotia: Lib 4, NDP 3, CPC 3, Green 1.
New Brunswick: CPC 4, Lib 4, NDP 2.
Québec: BQ 29Lib 18, CPC 17, NDP 9, Green 2.
Ontario: CPC 42, Lib 37, NDP 19, Green 8.
Manitoba: CPC 7, NDP 3, Lib 3Green 1.
Saskatchewan: CPC 8, NDP 4, Lib 2.
Alberta: CPC 20, NDP 3, Lib 3, Green 2.
British Columbia: CPC 17, NDP 9, Lib 7, Green 3.
Yukon: Lib 1.
Northwest Territories: NDP 1.
Nunavut: CPC 1.
Which gives us a grand total of:
CPC 122, Lib 85, NDP 55, BQ 29, Green 17.
Not radically different from the nation-wide D'Hondt: a few extra seats for the Conservatives and Liberals, a few fewer for the Greens - closer to reality, maybe, except for a further BQ drop. But you can see how having this additional distribution of numbers might make for a more convincing system: Prince Edward islanders, for example, would be able to identify the two Conservative MPs and the two Liberal MPs who were 'theirs'. The North, incidentally, remains exactly the same, since a prop-rep region of one seat is effectively a FPTP region.

Ontario and Québec in particular, however, remain pretty impersonal. How would someone from, say, Trinity-Spadina or someone from, say, Nickel Belt know which MP was 'theirs'? Which one to write a letter to for whatever reason it is that normal citizens write MPs? Sure, with 105 riding and 105 MPs, we could arbitrarily allocate one MP to each riding, but that would be a sheerly bureaucratic decision. Which eight ridings would get Green MPs?

The smaller a D'Hondt region gets, the less useful the D'Hondt system is. But if you broke up larger provinces into smaller regions, you might have results that gave an increased sense among the public that the MPs actually represented a constituency and not merely represented their parties. That's what I'd like to do, but it'll take a bit of time, and number-crunching...

Enhanced by Zemanta


  1. Why would you assign just one MP to one riding? If the Green Party has 8 MPs in Ontario, they will divide Ontario into 8 regions and assign an MP to each region. The CPC, with 47 MPs, will divide Ontario into 47 regions and assign one MP to each region. The Libs will divide Ontario into 37 regions, and the NDP will divide the province into 19 regions, and assign an MP to each region. The upshot is that every voter will have an MP from every party. Choose your representative. It's called democracy.

  2. Breaking up the larger provinces into smaller regions has already been done, by Prof. Henry Milner at Fair Vote Canada's electoral reform conference Feb. 21, 2009. You can see the results of the simulation here:

  3. (I'm shocked to see people even taking note of this tiny little blog. Welcome!)

    Wayne - that makes sense. It's a fair amount of map-manipulation, for each party each election to have to do, and I wonder how aware the average citizen would be of who represents them in such a case. But it's a good idea.

    Wilf, those are interesting numbers, and I like the smaller regions. I've been perusing them, but they're MMP, right? I was speaking about a 100% prop-rep system. I plan to periodically apply different models to the 2008 numbers, so it's still a useful resource and I thank you for it.

  4. Incidentally let me add a link for something I recently wrote that looks at some historical cases where FPTP has produced some curious results. It's more a primer for the non-Canadian, but it might interest:

  5. Jerry, MMP IS a 100% proportional system, if it has enough regional MPs. The model presented Feb. 21, 2009, had only 35% regional MPs, but that was enough in most regions. As compared with a perfectly proportional model, this model gave the Bloc three extra seats as a result of the low 35% ratio, at the cost of the Greens (two) and the Liberals (one). In the rest of Canada the 35% ratio cost the NDP two seats and the Greens two more, to the benefit of the Liberals (three, outweighing their Quebec loss) and the Conservatives (one). I'm certain the Greens would not complain; it's a minimal trade-off in return for a model that maintains local accountability and voter choice.