Electoral Reform

Some posts about electoral reform in Canada that I wrote in late 2015 and 2016. More recent posts can be found under the “electoral reform” tag.


  1. The Unintended Consequences of Proportional Representation
  2. The Preferential Ballot: Who Benefits?
  3. Proportional Representation: Who Benefits?
  4. Electoral Reform and By-Elections

The Unintended Consequences of Proportional Representation

Now that the election is over, many disappointed Canadians are talking about proportional representation, and how much fairer it would have been than our first-past-the-post electoral system.

I have never been a fan of proportional representation, partly because I think its supporters are mostly unhappy with the results of the most recent election and want to re-run it under different rules that they think are more favourable to their side, and partly because I’m a big fan of holding members of Parliament accountable to a constituency. But also because I’ve seen how PR systems operate in other countries. Not in the intricate, technical, here’s-how-it-works terms that PR advocates like geeking over — the real-world effects. And some of those real-world effects are precisely the opposite of what PR advocates want.

Let’s begin with party discipline, because there’s a lot of overlap between people who want electoral system reform and people who want their members of Parliament to be subject to less party discipline (in either case, they’re unhappy with the status quo). Problem is, PR and relaxed party discipline are mutually exclusive. If you’re elected from a party list in a pure or mixed PR system, your status as an elected representative is entirely due to the party and its fortunes. Make trouble, and you’ll be further down the list next time, if on the list at all.

There are, of course, options: if you have a mixed PR system you can still run as an independent, as you can in a FPTP system. But one option looms large in a PR system that isn’t nearly as workable under FPTP: you can split off and start your own party. In fact, a disgruntled representative has every incentive to do so, because you stand a much better chance of being elected at the head of the list of a smaller party that elects a couple of MPs than you do further down the list of a larger party that elects a few dozen.

In a FPTP system, parties tend to be coalitions of like-minded but rather discrete factions that are often at odds. Social conservatives and fiscal conservatives don’t always see eye to eye inside the Conservative Party, the Liberals’ focus on civil liberties hasn’t always sat well with its historically Catholic voting base (abortion, gay marriage), and the NDP is often a battleground between its various constituencies (case in point: the Waffle). But these factions hang together out of mutual self-interest, because failing to do so guarantees a lack of electoral success.

But in a PR system, these internal factions would very quickly become external. Each of Canada’s major parties would split into several smaller parties, because it would be in their self-interest to do so. For one simple reason: leverage.

Take the Conservatives. Say what you will about our dear, soon-to-be-departed Stephen Harper, but one thing he did was refuse to reopen debates on things like abortion, capital punishment and same-sex marriage that social conservatives in his party would dearly loved to have had. This was electorally expedient for him: in order to take power he needed the votes of moderate Canadians who would have been been repelled by a such an agenda.

But in a PR system the social conservatives would almost certainly have their own party. Instead of being told behind caucus doors to bite their tongues, they’d be able to advance their agenda by making it the price of their support for a coalition government.

Because almost every parliament after the implementation of proportional representation will have a coalition government. That’s good, PR advocates say: governments will have to draw their support from several parties with diverse views and their policies will represent a greater swath of public opinion. Right?

Not necessarily.

It’s not necessarily going to be like Germany, with its mixed system and five-percent rule to keep fringe parties out, where large, moderate parties enter into grand coalitions with one another. Sometimes it’s like Israel, a pure PR system where Benjamin Netanyahu won only 23.4 percent of the vote (and 30 out of 120 seats), but was able to form a coalition with the help of Orthodox religious parties.

Sometimes coalitions draw their support from the extremes, not the centre.

So if Canada were to adopt some form of proportional representation, the election results wouldn’t automatically lead to a minority Liberal government obliged to enter into a coalition with the NDP. Because there would be no advantage for Canada’s political parties to remain as they are under proportional representation. Instead we might see jockeying among a dozen or more parties — prairie-based Christian conservatives, urban radical socialists, right-wing Quebec nationalists, Charter-obsessed liberal democrats, Ontario rural conservatives, an actual labour party — for position and advantage.

The resulting coalition might well be less moderate or inclusive — or even representative! — than a majority government would have been under FPTP.

It woudn’t necessarily be worse than what we have now, so long as you understand that the price of having a successful Green Party with a dozen seats is probably a Christian conservative party with forty seats. It does cut both ways.

I just think that PR advocates haven’t considered what the unintended consequences might be if they actually got what they wanted.

October 20, 2015


The Preferential Ballot: Who Benefits?

Of the electoral reform proposals on offer, the one I’m most in favour of is the preferential or ranked ballot, also known as the alternative vote, where voters mark their first, second and third (and so on) choices (rather than simply an X). If no candidate wins 50 percent on the first round, the candidate finishing last is dropped and their voters’ second preferences are redistributed, and then the next candidate, and so on and so on until someone gets 50 percent.

It’s a bit more complicated than marking an X, and counting (and calculating!) the votes will almost certainly take longer, but a preferential ballot has a couple of things going for it. For one, it ensures that any member of Parliament will eventually be elected by 50 percent of the vote, once second- and third-place preferences are redistributed — we won’t have close three- or four-way races where the winner ends up with only a handful of votes more than the others, and less than 40 percent support.

For another, it’ll make strategic voting obsolete. No longer will I have to wonder whether voting for candidate X instead of candidate Y will allow the foul candidate Z to get elected: I can simply rank candidates X and Y on the ballot and leave candidate Z off the ballot altogether. (I have strong objections to strategic voting: in practice it’s been weaponized to get other parties’ supporters to support your candidate, and parties haven’t always been honest about who’s ahead and who’s behind when doing it.)

But changing the rules does benefit some players more than others. This is true of any change, including proportional representation, which is one reason I suspect electoral reform has been so elusive: the sense that the rules are being tweaked for someone else’s benefit.

In the case of the preferential ballot, that someone else is largely expected to be the Liberals, because as the centrist party they’re most likely to get second-preference votes from both the right and the left. No surprise, then, that a preferential ballot is reportedly the Liberals’ preferred option.

It’s one thing to expect something, quite another to quantify it. What would the 2015 election results have looked like with a preferential ballot? Two separate models have tried to answer that question, each using slightly different methodologies: Éric Grenier of ThreeHundredEight.com and University of Lethbridge political scientists Harold Jansen and Peter McCormick. The following table shows the results of their models, along with the actual 2015 results and what the results might have been under proportional representation:

Party 2015 Results Grenier Jansen/McCormick Proportional
Liberal 184 224 205 134
Conservative 99 61 73 109
NDP 44 50 52 67
BQ 10 2 7 16
Green 1 1 1 12
Preferential Ballot Models, 2015 Canadian Federal Election Results

While necessarily inexact — the models apply regional polling data on voters’ second preferences to individual riding results, and the impact of strategic voting, which would be different under a preferential ballot, is impossible to quantify — it can fairly be said that a preferential ballot would largely benefit the Liberals (and to a lesser extent the NDP), at the expense of the Bloc Québécois and the Conservatives.

A preferential ballot benefits both the Liberals and the NDP because many of each party’s supporters will opt for the other one in the next round. And it punishes the Conservatives — but mostly because they’ve alienated just about everyone except their core supporters. Polls have repeatedly shown that few Liberal and NDP voters have the Conservatives as their second choice, and many Conservative voters say they don’t have a second choice.

That might change under a preferential ballot. Grenier argues that parties would change the way they campaign under a preferential ballot — they’d have to: “Energizing the base and taking swing voters away from other parties would no longer be enough — leaders would also need to woo supporters of their opponents for those second-choice votes, and avoid alienating other voters entirely.” So the Conservatives would have every incentive to be nicer to people outside their narrowly identified base. And so would every other party. Because if they didn’t, they would lose every close riding once the second and third preferences were redistributed.

It would not change the system very much at a macro level: Jansen and McCormick calculate that only 32 seats would change hands, and the system’s tendency towards majority governments would be even stronger. But at a more granular level some of the more toxic tendencies of our present system — narrow wins with a split electorate, strategic voting, riling up the base instead of appealing to the broad electorate — might be addressed.

December 2, 2015


Proportional Representation: Who Benefits?

In my post on electoral reform and the preferential ballot, I remarked that changing the rules benefits some players more than others. This is true of any change, and I suspect it’s one reason why electoral reform has been so difficult: the sense that the rules are being tweaked for someone else’s benefit.

As the centrist party, the Liberals would probably benefit from ranked or preferential ballots (and modelling appears to bear that out), and, funnily enough, ranked or preferential ballots happen to be Prime Minister Trudeau’s preferred option. So it’s easy to complain that the Liberals are rigging the game in their own favour.

But it’s important to keep in mind that, despite the lofty rhetoric about a “fairer” electoral system in which “every vote counts,” proportional representation also benefits some players more than others. The question is, which players?

Political parties for one: that much is obvious. (Although, as I’ve pointed out, not necessarily the current political parties. There would be a realignment.) But which political parties?

To answer that question, I looked at the results of every Canadian federal election going back to 1980, and compared what the actual results were under first-past-the-post with what they would have been under pure PR.

Now before we continue, we have to be mindful of a couple of things. For one, no one is actually proposing pure PR: most electoral reformers propose mixed-member PR or a single transferable vote. But we tend to use pure PR as our example because it’s a hell of a lot easier. For another, the electorate would almost certainly vote differently under PR or any other major rule change. There are too many variables at play, especially for someone as stats-challenged as myself.

But with all that in mind, what might PR have done, historically? Here’s a graph that shows the difference — not by which party, but the place in which a party finishes (i.e., in 2015 it was the Liberals, but in 2011 it was the Conservatives).

Difference in results between FPTP and PR in Canadian federal elections, 1980-2015
Difference in results between FPTP and PR in Canadian federal elections, 1980-2015

On average, the government party would lose 39 seats under PR. The official opposition party would gain two or three seats on average — basically a wash. The prime beneficiaries are the parties in third, fourth or fifth place. How much do they benefit? Here the Bloc Québécois confuses things, because their vote is concentrated in one province: they actually benefit from first past the post, and lose seats under PR — an average of nine. For other parties in third, fourth or fifth place, the average gain is around 22 seats.

So, compared with first past the post, PR is basically a means of transferring seats to the smaller parties. Not for nothing does PR tend to be popular among the parties that often finish in third, fourth or fifth place. They’re advocating for their own self-interest.

But what about the fringe parties? Under a pure PR system, some of them might actually have elected an MP. The Greens might have elected at least one MP since 1984; they’d have elected 21 in 2008 instead of none. But we’d also have seen MPs from the Christian Heritage Party, Libertarian Party, Marijuana Party, National Party — and, in 1993, two MPs from the yogic-flying Natural Law Party.

You have to admit, it would have made Parliament more interesting.

But if Canada were to adopt a system whereby political parties needed a minimum percentage of the popular vote to get party list representation in the House of Commons (as is the case in Germany, where the threshold is five percent), all of these parties would be shut out. As might some current parties: at the national level, neither the Bloc nor the Greens got more than five percent of the vote, though the Bloc got 19 percent in Quebec and the Greens eight percent in B.C. Under a mixed system they’d certainly elect individual MPs, and under a regional party list system they’d do all right — but almost every system short of pure PR would keep out most other fringe parties.

Unless, of course, electoral reform changes the behaviour of the electorate. Which it almost certainly will: voters would have almost no reason to vote strategically, and would be able to vote for their first preference, no matter how fringe or nutty. Mainstream coalition parties would fragment to cater to those preferences.

And there’s every chance of Canadians taking the opportunity to prank the system. We’ve got a nihilist streak in this country, one that tried to change the name of the Northwest Territories to Bob. One that gave the Rhinoceros Party several second-place finishes in the 1980s (under pure PR, the Rhinos would have elected three MPs in 1980). One that gave 84,743 votes to the Natural Law Party in 1993 precisely because they were batshit bonkers.

Just imagine what we could do if a few pranksters spread out across the country could get somebody elected. The political class clutching their pearls will only encourage us. Try to stop it with a five-percent rule and we’ll take that as a challenge. You just watch us.

That might actually be the best argument in favour of proportional representation, now that I think of it.

January 15, 2016


Electoral Reform and By-Elections

In a piece discussing four upcoming by-electionsiPolitics’s Susan Delacourt mentioned in passing that “unless electoral reform happens really quickly, those seats will be filled the old-fashioned way — no preferential ballots, winner take all.”

Which made me wonder: how exactly would electoral reform deal with vacancies in Parliament? By-elections, after all, only work if MPs are elected from individual constituencies; you can’t run a campaign across an entire province or country just to fill one or two seats out of more than three hundred.

I could speculate, but instead I had a look at the website of Fair Vote Canada, a group advocating for proportional representation in Canada, to see what their solution would be. In their submission to the Special Parliamentary Committee on Electoral Reform, FVC proposed three options, one of which was mixed-member proportional representation, in which regional seats, elected from party lists, are added to MPs directly elected by their constitutencies. In that option, regional seat vacancies would be filled from party lists.

In case of a resignation or death of a regional MP during a term, the party’s runner-up moves up into the seat. No working MMP model has by-elections for regional MPs. As the Jenkins Commission pointed out, if a region-wide contest were to take place “it would almost by definition result in the victory of the predominant party in the area, thus negating the essential purpose of the Top-up seats.”

I have to say, this isn’t something that makes me more likely to support proportional representation.

August 24, 2016