Tuesday, February 5, 2013

A Primer on AV vs PR

There's some controversy these days surrounding most of the Liberal leadership candidate's penchant for supporting the "Alternative Vote," or preferential ballot vote (AV from now on), to replace our current first-past-the-post (FPTP) voting. I believe eight of the nine candidates have expressed support for changing the way we elect representatives, while Joyce Murray seems to not have actually declared her preference but instead leaving it open to some form of proportional representation (PR).

In fact, there has even been a whole group and website dedicated to pressuring the candidates to support PR over AV, seemingly backed by Fair Vote Canada. What form of PR they want - whether its the MMP system suggested in the past, or pure PR like the Israelis have, or something some obscure European country uses - is unknown. I doubt its MMP, simply because that has a mixture of FPTP constituencies (which these people despise) and PR.

Anyways, they describe AV as simply a lateral move from FPTP - that AV will produce little to no change in the system, and that Canadians will not have true PR.

Their arguments, as far as I gather, break down thusly:

1. AV is good only for single-post positions, like for mayoralty or presidency, but not for legislative bodies (councils, HoC, etc.), as it is still non-representative of voters

2. Like FPTP, not every vote will count in an AV election, as there will only be one representative elected, with the supporters of the losing candidate not represented either

3. AV builds strategic voting into a system, instead of eliminating the need for it

4. 70% of Canadians want some form of PR instituted

I may be missing a couple of arguments, but these are the ones that I found basically being the most important to these folks. The basic, overriding theme of their campaign is that any move to change FPTP should be towards making the way we elect MPs and so on more representative of how people voted as a whole.

The best way to describe what they want, in very simple terms, is in graphical form.

For the 2011 results, under pure PR, and assuming they're proposing that parties must reach a certain limit (like, parties must reach 2% of the total national vote to actually have seats, or however you want to apportion that), the Conservatives would be limited to a small minority government, and the Liberals and NDP with 154 seats. Add in either the Bloc or the Greens, and an easy "progressive" majority, or whatever they're calling it these days.

Its utterly simple, in this case - the percentage of votes earned nationally equals the percentage of seats a party gets. I was going to do this based on the provinces as well, but decided against it because of the fact that the amount of seats each province gets in the HoC is not proportional to their populations, thus making wonky seat totals. I assume that's going to have to change too.

Again, this is simply pure PR. If you were to go to an MMP system, like what was proposed awhile back in Ontario and some other provinces (and soundly defeated), things would look a tad different. To explain MMP very simply, you elect your candidates in these single-member constituencies like we have now, and the "list" candidates are apportioned based a party's vote percentage-to-seat percentage gap. For example, in 2011 the Conservatives won roughly 54% of the seats on 39.6% of the vote; assuming there are 100 "list" MPs to be elected, the Conservatives would likely get none. The Liberals, who won 11% of the seats based on 18.9% of the vote, would need some of these "list" MPs added so their total seat count somewhat ended up matching their total vote percentage.

Again, this is maybe best represented in graphical form:

This is a very simplified version of MMP, as there are a lot of nuances to get into - but, basically, you can see what happens here. It isn't necessarily always exact, as New Zealand's elections should be an example of (the National Party in 2011 ended up with 48-49% of seats, on 47% of the votes), but it tries its hardest.

There are a multitude of other ways to get people elected, but these versions come up the most often, and are what I assume people at Fair Vote and "Liberals for Fair Voting" have foremost in their minds. I could be wrong, but again, they don't actually say what a better alternative would be, just that AV is bad.

So, what about AV? I think most people understand what it entails by now - you rank your preferred candidates in order, and as one by one, candidates are eliminated, their second, third, fourth, etc., preferences are apportioned to remaining candidates. Awhile back, I did this in my own riding of Burlington and the Quebec riding of Gatineau specifically, then did rough calculations across all of Canada - but based on the 2008 results. I'm unsure now how I got those numbers, but hey, I have faith in my younger self.

Its hard to say how an AV Canada would look like really, but the point of it would be that every single riding of Canada would have a candidate that would win over 50% support of the riding's voters. The idea is that no longer would candidates win on something as low as 31% support, like Hedy Fry in Vancouver Centre; instead, the winning candidate would need 50%+1 support total to become the riding's representative.

However, as these PR pushers have said, this isn't actually proportional. All it guarantees is that representatives have earned majority support within their constituencies. You could, feasibly, have the AV system allow a majority government with the Conservative's 39.6%, so long as they managed to get over 50% support in 155 ridings.

However, this is more difficult than it sounds. There were 107 Conservative candidates that won over 50% support in 2011, with a further 28 winning between 45-50% (a fairly guaranteed-to-win number that is below the threshold). That gives the Conservatives about 135 seats they can rely on to win under AV, and earn roughly 40-45% nation-wide in terms of support. In order to each the 155 seats needed for a majority, the Conservatives will need to be fairly close to, if not above, 45% support across the country.

While it obviously isn't proportional in an exact way, it does mean that the Conservatives will either need higher initial vote tallies, or enough transfers from defeated Opposition candidates to win. No matter what, the Conservatives, or any party, cannot sit below 40% and win a majority government, even under AV. The ways to do it are just so extreme as to be a waste of time.

The problem will come in for smaller parties - specifically, in 2011's case, the Liberals. We won only 7 seats with 45% or more in 2011, and 15 with 40-45% support. Under AV, it could be incredibly harder for us to win seats with 50% or more. That is just a fact.

The key thing to remember, though, is that the Conservatives are probably going to earn less support from defeated NDP candidates, than they will from defeated Liberal candidates. Thus, in Mississauga, where a lot of those ridings were in the 35-40% range, and the NDP sat between 15-25%, we could see the Liberals win those kinds of ridings based on NDP transfers, even if they're not necessarily in our range - in fact, most would be in the Conservative's range.

The issue is, we really don't know how these transfers would break down. It is entirely possible that the Conservative ridings with 45-50% initial support, but facing strong opposition from a Liberal or NDP candidate (say between 35-40%), could lose their seat on the back of transfers between these two Opposition parties. It is a risk, but it's one that could pay off.

Like the PR folks say, it builds in strategic voting. It ensures a constituency is represented by the candidate that a majority of voters decided to back in some way. It is not proportional representation, though it does require parties to increase their vote totals, one way or another, to earn a majority.

If you're looking for the straight battle between which is more representative on a grand scale, PR will win, hands down.

But, if you're like me, then keep reading.

My question is not over the proportional nature of AV vs PR - I know which won wins. The thing that is unsaid from all these PR advocates is that by instituting PR, we're going to end constituency representation.

Pure PR has no need for local constituencies in the current way we think of it. PR constituencies are often regional or even national, and they cover wide swaths of territory. The representatives elected come from lists, and are therefore indirectly elected representatives, because really the voters have no true say in which list candidate becomes elected. The list positions (1st, 2nd, 3rd, etc.), or the order in which list candidates are elected depending upon the seat count (i.e., the 42nd candidate on the Liberal list won't get elected if the Liberals don't have enough votes for 42 seats), are usually chosen in primaries or appointed by the party leadership. We have no choice if Barry Biblethumper, the 104th Conservative list candidate, is elected to represent us - that's just how it will end up.

MMP offers a compromise in a way, though its generally half-assed. AV offers a compromise as well, except its a bit different; you can keep your local constituencies, and directly elect your representative, but that candidate requires majority support, so no one can say they were elected unfairly.

I like that compromise, though I can handle MMP's as well. I prefer having a local representative for my constituency, rather than some list candidate I won't be directly electing anyways.

But, that's just me. I wanted to give a bit of a primer for anyone interested in this stuff, because it could become a bigger issue down the road.


  1. The only form of Proportional Representation that'll work in Canada, is a Parallel system like they use in Japan.
    This is a picture of how that'd work. Note the "list seats" and "votes" are equal share.

    1. Hm, interesting stuff Teddy. I didn't think about that.

  2. A great writeup, Kyle. I think you may have understated the worst aspect of PR, though: ridiculous, single-issue fringe parties holding the entire country hostage to their ill-conceived demands. I look at all those European basket-cases that now teeter on the brink of bankruptcy, and I see PR present in every one of them. I want no tree-huggers, yogic flyers, ex-porn stars or religious fanatics holding the balance of power in our government; the Natural Law Party, the Sex Party, the Marijuana Party and the Rhinoceros Party may be fun to think about, but they would be a nightmare if actually elected to Parliament.

    It's AV all the way, for me (or run-off voting, if you like).

    1. Thanks Fred. You are right, I did forget to mention this, and it is a huge concern for PR - I mean, just look at Israel. The difference being, however, that I don't see many of these fringe parties gaining traction here in Canada - and even if they did gain some, the key thing is to have a proper threshold. I personally think 5% is good, based on provincial vote counts; that way, at least the Greens get some seats, and it ensures that you can't start some little spin-off party and gain power. I believe this is what they've done in New Zealand, they've got a pretty high threshold and it works out well for them.

      But, really, who knows what could happen? Like you said, it could be dangerous to allow personality parties and single-issue groups to hold the balance of power in Parliament. Instability is not something I support.

  3. I choose AV as well - for the simple reason PR and every form of it has been soundly defeated in referendum. AV has been used by all the major parties to elect their leaders.. so it's not alien to them.. and it can be explained simply to the electorate. It is also the least "radical" of the voting reforms out there.. and Canada has shown it is conservative when it comes to voting changes.

    My support for AV over a form of PR isn't because I think it's better necessarily.. but because it's the most feasible and practical option to push at this time.

    1. Good points, Scott. That is also true, though I would point to the UK and BC as examples that maybe even AV isn't going to gain traction, especially if it goes to the people. You are right, though - this should be acceptable to all parties, and thus the easiest reform to get out there.