British Columbia rejects PR


The current first-past-the-post system received 61.3 per cent of the votes cast in a mail-in referendum, while proportional representation got 38.7 per cent in the results released Thursday by Elections BC.


Well it didn’t help that voters didn’t even know what form of PR they were voting for since the PR voting method was to be decided in that same election. It would of been smart to hold 2 separate referendums…


Do we also have the results for what method would have won had BC voters opted to change the voting method?


Or, if you’re going to use IRV to pick the PR method, why not just make “Plurality Voting” one of the options?!?! Argh.

In some sense this might be a wake up call though. The FairVote crowd might want to pause and consider what’s more achievable.

This is in line with the Montgomery County (Maryland) thing. If it passes, then the county council gets to decide whether to adopt IRV or Approval Voting. What do you think they’ll pick given the option? The more complicated one? I don’t think so.


And allow non-monotonicity to effect the PR vs Non-PR election result? Even if IRV was monotonic, if plurality voting happened to be the Condorcet winner and IRV did not pick it, that would infuriate the non-pr voters. Perhaps if they were going to use a ranked ballot, they should of just picked the voting method that pairwise defeated the current voting method by the greatest margin, and if none did, then default to the current voting method. That’s a clever way of squeezing in approval voting by making plurality voting the threshold in the case that one of the PR methods wins.


My 2 cents:

(1) The biggest problem was that the Attorney General decided the design of the ballot and the PR proposals single-handedly and without any public discussion. There would have been more public interest if these questions had been decided by the parliament, a parliamentary committee or some other public commission.

(2) The PR proposals were too similar and too vaguely defined. There would have been more public interest if the voters had been able to choose from a larger number of more diverse proposals.

(3) The Green Party of Canada and FairVote Canada should rethink whether it is a useful strategy to refuse even to discuss single-winner reform.


This strikes me as extremely implausible. No one understands PR methodologies and their intricacies. I couldn’t even get San Francisco tech workers to explain IRV correctly. It would have been better to have just a single option vs. the status quo. A simple choice.

A single-winner reform, like Approval Voting, strikes me as far more likely to succeed. If it can pass in Fargo by 64%, I think that shows much greater political viability than complex PR schemes.


Is there something uniquely difficult about passing voting reform in Fargo?


I live in British Columbia so actually got to vote in the referendum. Additionally, I submitted a proposal into the to The Attorney General along with many others. I was surprised to find that most of the other submissions did not express support for a specific system. I endorsed Approval Voting with a change to a kind of Consensus Government formation. I am surprised how little talk there is about government formation among election reformers given that one of the major reasons for wanting Proportional Representation is because of how government formation is partisan.

Anyway, I was pretty disappointed that they did not even mention Approval Voting in their review document of all the submissions. Further disappointed in the choices they gave. In the end my ranking/scoring would have been:

  1. Dual Member Proportional (score = 9)
  2. Single Member Plurality (score = 8)
  3. Mixed Member Proportional (score = 2)
  4. Rural Urban Proportional (score = 0)

Given that the polls showed that MMP was going to be the favourite (and it was) I ended up voting no to PR. Maybe FairVote would rethink their strategy based on this. A single score vote just seems so obvious for this circumstance.

There will be similar referendums in PEI and Quebec in the upcoming year. I think they both plan to put forward MMP and have no public consultation.


And “keep plurality voting” should have just been an equivalent option, amirite?


Why SMP above MMP?

When voters are maximumly strategic, it is theoretically possible for MMP to reduce to parallel voting with SMP + some party list seats, so do you not think PL is better then SMP?


I think Party List is by far the worst system which could be considered a representative election. Ill get back to why.

For MMP vs SMP it is sort of tricky given that they did not specify which variant of MMP. But lets just assume parallel voting for simplicity.

Since MMP is really just SMP + PL you get the issues of both and don’t actually solve vote splitting but make up for it somewhat by getting full Proportional Representation. But that proportional representation comes at a cost of other representation.

Proportional Representation

Local Prepresentation
Balanced representation (That each member represents the same number of citizens thus legitimating parlimentary votes)

I would say that the overall representation of the diversity of opinion would be a net increase at this point BUT… since there is a PL component there are partisan votes. This has the consequence of an increase in party power and hence partisanship in the system. Furthermore, the ideological rifts inside parties tend split them into smaller parties which are more ideologically homogeneous. The power coming from the top to give candidates who are more ideologically homogeneous leaves a lot of spaces in between parties and hence gaps in coverage of the space of ideological opinion. This results in an issue where the whole of the ideological center who hold nuanced ideas and partially supports many parties on specific issues are not expected to be represented by any party and hence less likely to be represented by anybody in parliament than in SMP. In his landmark book “The Logic of Collective Action”, Mancur Olson referred to these people as “The ‘Forgetten Groups’, Those Who Suffer In Silence”. There are many examples of this issue happening with tax codes. Hopefully that made sense, it is hard to summarize why partisan votes are a bad way to get Proportional Representation succinctly. Essentially, Partisan votes ruin the metric of Proportional Representation for evaluating non-partisan but ideological representation.



Worse then FPTP?!?



Have never even heard anybody argue otherwise except in small countries like Israel where geographical differences are not important.


Here’s some quick notes on my perspective here.

First, some background.

Basically, this referendum was mishandled from the start. When the NDP/Green government formed, they put the provincial attorney general (Dave Eby) in charge of structuring the referendum. Instead of emulating the successful 2004 Citizens’ Assembly (which led to PR getting 57% support in the 2005 referendum), they made a sad copy of Trudeau’s sweep-it-under-the-rug public survey, with the same false dichotomies (“Do you want wasted votes, or do you want hopeless gridlock?”)

That prompted the CES (that is, mostly, me, with help from Mira Bernstein) to organize the, so that there would be some real expert input. The symposium was helpful, but we didn’t feel we had a clear legitimate mandate to push to a single consensus recommendation, so we mostly confined our recommendations to listing several serious PR options and best practices around each of them.

Most of our recommendations were influential in the design of the referendum, with two big exceptions. First, we recommended that if MMP was an option, it should be clearly specified that it would be open list, based on the Bavarian model. In the referendum, MMP was left unspecified; then, after the start of voting, leaders of all 3 major parties stated that they would push for open list if MMP won. Thus, the lack of official specificity there was completely pointless. Second, we recommended that there should be a quick citizens’ jury to give nonpartisan consensus recommendations to voters. There was (barely) time to do this if they’d hurried; but they didn’t. Thus voters had nothing but the media (largely biased against PR), or partisan/activist sources (less credible than nonpartisan, even when good).

The structure of the referendum had all the fingerprints of designed-by-committee-with-insufficient-incentive-for-consensus, but other than that, was not in my opinion as bad as has been portrayed in this thread. Yes, certainly, if it had been just me designing it, I would not have done that two-question majority-then-IRV structure. But there is some logic there.

  • Having the plurality-vs-PR question separately, up front, gives voters an easy way to opine there without wading into the details of the different PR methods.

  • There are obviously better methods than IRV for the second question, but the difference they’d make is minor, given that there’s no obvious center squeeze situation between the three options. IRV is the system more people have heard of. I definitely pushed against that but at a certain point you have to go with the flow if you don’t want to turn people against you.

  • Having plurality not be an option in the second question, given that it was already there in the first, was seen as a way to simplify. I disagree but there is some logic there.

The campaign was also mishandled. The NDP support was very weak, with some NDP politicians actively undermining the PR campaign. The “PR is lit” messaging, an attempt to get youth involved, may or may not have worked with its target audience (self-conscious irony winning out over cringeworthy “how do you do fellow kids” pandering?) but almost certainly was a huge turnoff to older voters. Most newspapers were pretty unabashedly anti-PR. Very poor outreach to the Chinese-speaking community. The official “yes” campaign was very bad at mobilizing grassroots energy, so there were various individuals who did more with nothing than the official campaign did with a half million dollars.

I could go on and on, but my tld̦r conclusion here is that the biggest, most basic mistake was not including any aspects of deliberative democracy (citizens’ assembly, citizens’ jury, or deliberative poll). The final outcome was 58% “no”, but that was actually substantially fewer voters than the 57% “yes” in 2005 after the citizens’ assembly.


Thanks for the insight Jameson. I had a look at the article but did not see any arguments for MMP. I understand that the situation already had you constrained to exact PR models so Approval and Score would not be accepted. This is a shame since there was bipartisan support and the level of PR would increase.

Anyway, are you arguing that Bavarian MMP is the best system or just that it was the most likely to win? I would think that DMP is almost clearly better. I personally would veto any system with explicitly partisan votes but it was not stated if they were an option or not.

What about RRV or SPAV? Where they even brought up at the symposium.


I wasn’t saying that Bavarian MMP is best overall, just that it’s the best form of MMP (as well as being the MMP version that would be most likely to pass in BC).

That’s especially true for modified Bavarian MMP, where the local and regional votes are treated for the purposes of party ratios almost as if they were first and second choice in STV. That is, if your local vote wins, your regional vote can be “used up” to fill the quota; and if just one of the two vote goes to a party which gets no seats, it transfers to the party of your other vote.

Personally, I also think that Bavarian MMP is also better than DMP, though this is more debatable than its superiority over say New Zealand MMP. DMP is essentially biproportional MMP; that is, where the top-up seats are constrained to be one-per-district. As such, it’s inferior to Bavarian MMP in that it has less voter choice; you cannot vote for a party without voting for the local candidate of that party, and you have no within-party choice.

As for RRV and SPAV, I think these are both very problematic regarding strategy. Allocated score voting, allocated STAR voting, allocated MJ, or some other allocated system would be better than any of the options that were on the referendum; but these were considered too out-there for the symposium to advocate. (Whether you use score or STAR or MJ or 3-2-1 or any other rated-ballot system under the allocation layer is almost immaterial; all such systems would get very similar results. Note that STV is essentially, though not exactly, the same as allocated IRV.)


Not really sure what you mean by allocated voting. Surely you can’t mean this.


Allocated proportional voting is an iterative algorithm that can be combined with any single-winner method. To fill each seat, first use the single-winner method to find the winner, then “exhaust” (that is, stop counting in future rounds) the “quota” of ballots that “most” support that winner. The meaning of “exhaust” is unambiguous, but there are variant allocated methods with slightly different interpretations of “quota” (probably Droop or Hare) and “most” (different sort algorithms that account for ratings for non-winning candidates differently; though generally, ballots with lower ratings for non-winning candidates are exhausted first).


Ahh OK. This is actually very similar to something I came up with as a replacement of RRV’s reweighting. Please comment in that thread.