BC: Sticking with FPTP

I thought it would at least be close. It was not.

61.3% of voters in British Columbia voted to keep the FPTP system.

In the (now moot) second ballot question, MMP was favored over the other options, on 41.2%, with DMP a (surprising?) second on 29.5% and RUP on 29.3%. A total of 831,760 votes were cast in the second question, versus 1,380,753 in the main question. So fully 40% of voters did not answer the “what if” question.

The second question permitted a ranking of choices on the three alternatives. In the redistribution of preferences, MMP won with 63.05% (of 779,698 votes).

In 2009, I wrote a post entitled “BC: FPTP forever?” And then the question of change came up again a mere nine years later. But this is a pretty decisive defeat. I know forever is a long time. But it seems unlikely for proportional representation to get another chance anytime soon.

52 thoughts on “BC: Sticking with FPTP

  1. The correlation between the percentage of votes for FPTP in a district and the percentage of votes for the Liberal candidate in the 2017 general election is R=+0.83. The FPTP-NDP correlation is R=-0.57 and the FPTP-Green correlation is R=-0.52.

    • One of the traditional arguments in favour of FPTP has been district size. If it was a factor in the 2018 PR referendum, it did not correlate with district size in a simple way.

      I removed the observed correlation between percentage vote for FPTP and vote for Liberal candidates (assumed linear relationship) and compared adjusted FPTP vote to log(district area) for the 87 districts. The correlation was R=-0.21.

      Is log(area) too crude a proxy for difficulty of servicing a district? is there some other, confounding variable? Or was district size truly not an issue for the PR models on offer?

      • Dave, this is really interesting. (I’d love to play with the dataset if you’d consider sharing.)

        I think log(area) should be an excellent proxy for the challenges of district service. Perhaps a more direct one is population density. Obviously, that would be strongly correlated to size, although perhaps not completely so, given that the rural districts tend to have lower population.

        If district area/density was not that much of an issue, it would be credit to the design of the systems on offer, as each had its own way of attempting to correct for the perceived size problem in the STV map of 2009. So the reformers could take some solace from that, I guess.

      • MSS
        Yes, I’m willing to share the BC 2017 election/2018 referendum dataset. (How?)

  2. I wouldn’t say forever, but at least until there are another two consecutive provincial elections perceived as FPTP failures, as were the 1996 and 2001 votes which led to majority support for PR in 2005 (but a near miss due to the impossibly high 60% threshold). The two consecutive perceived failures seems to be the key for getting the ball rolling, judging from past developments not only in BC but also in New Zealand (electoral reform triggered by back-to-back majority reversals in 1978 and 1981) and the U.S. state of Maine (a controversial, polarizing figure elected as governor on a plurality in 2010, and re-elected in 2014 on a second plurality).

    In BC’s case, FPTP backers made much of the claim that PR would lead to the entry of extremist parties in the provincial legislature, ignoring the fact that under FPTP a major party could gain power on a minority of votes after having been hijacked by an extremist faction – and one doesn’t have to go too far from the province to find that very scenario already in play. I wouldn’t be surprised if in an ironic turn of events that also comes to pass in the province in the years to come, in due course forcing the electorate to reconsider once more its position on the matter.

    I didn’t focus too much on the proposed alternative systems because I’ve been busy with other matters (among them Maine’s successful general election rollout of AV, or RCV as it’s called there) and also because I was of the view that it was quite possible BC voters would reject change, which turned out to be the case. That said, I thought as well it would be a close call.

    • I would say not forever, but certainly not before PR gets adopted and used once or twice in other parts or levels of government in Canada. So despite BC having tried to pass PR more than any other part of Canada, it most likely will only adopt PR once it has made its entry elsewhere.

      • Yeah, my general view is that this is an issue that will not go away. The inherent conditions are there and are likely remain there. But it may take a series of further contingencies (such as Manuel notes) to turn things in favor of PR. I do not know if adoption of PR elsewhere in Canada could be one of those contingencies, but it certainly is plausible that such reforms could help.

    • There’s also Quebec’s recently-elected CAQ government’s pledge to implement within a year MMP for National Assembly elections in the province, but given that the existing FPTP system gave them an unexpectedly substantial seat majority earlier this year, it remains to be seen if they’ll develop a case of cold feet on the issue.

    • Even PEI looks like a heavy lift. For MMP to replace FPTP, it must have a majority of votes in at least 60% of the 27 districts. That is, change requires carrying at least 17 districts while the status quo needs only 11 districts.

      Suppose voters in each district were divided 50:50 so that it was a matter of chance whether FPTP or MMP won a given district. Then, the probability of winning 17 or more districts would be 12.4%.
      Another point of reference: in the 2016 plebiscite a 52:42 advantage in final preferences allowed MMP to carry 22 districts. So maybe 6–8% spread needed for reform?

      • Slightly off topic, but MMP is a weird solution for a place as tiny as PEI. Nowhere would the case for STV be stronger.

      • Why MMP? The 2015 election gave the Greens only 10.8%, and the NDP only 11.0%. The 2011 election gave them 4.4% and 3.2%, and so on. Back in 2000 the NDP got 8.2%. To make every vote count in PEI, with 27 MLAs they need province-wide proportionality.

      • Even FPTP hasn’t stopped the Greens from winning a seat, so they’d have no problem winning seats under STV, and I think NDP could also easily win a seat in some 5 or 6-seater.

      • I did not know that they had put in the provision requiring 60% of districts. That may well doom it (as presumably intended). Nothing like stacking the deck in favor of FPTP by making a referendum contingent on success within a super-majority of existing districts!

        As to the question of MMP vs. STV or other system, it is quite striking that STV was not even one of the options in the earlier mail ballot.

        However, I don’t think size should argue against MMP. If anything, it makes the case stronger, as the jurisdiction-wide compensation district is not massive as it is in New Zealand, or as it would be in BC if a non-regional MMP model had been on the table.

        Having said that, I agree that STV would make a lot of sense for PEI. So perhaps would DMP, given that they had a form of two-seat districts in the past that such a proposal builds on.

      • Small size of the jurisdiction is an argument in favour of STV simply because one of the main arguments against STV (at least in BC) is that it would make districts too big.

      • Sooner or later there will be a Canadian ER referendum where the requirement for change is a majority in 1 more than the number of electoral districts that actually exist.

  3. No two reverse pluralities in a row would cause a change in the electoral system. It is funny that FPTP jurisdictions that want to move to PR are talked about and consistently in the air, and NZ is one of those that changed back in 1996, but no PR jurisdiction ever wants to go to FPTP system. What is the reason for that?

    • Italy (sort of), Japan (sort of), NSW, NYC, Isle of Man all went from PR systems to non-PR or semi-PR systems.
      Even Germany’s 1949 move from pure lists to MMP was viewed as a partial step back from “pure” PR, inasmuch as the districts can change the outcome from what nationwide or State-wide closed lists for 660 seats would produce.
      But yes, once voters have tried a preferential and/or proportional system, it is usually – outside of Canada with its “Pirates of the Caribbean”-era Constitution (I mean, where else are Senators “summoned” to their seats?) – hard to persuade them to revert to FPTP. Just as secular jurisdictions with a 365.2525-day solar calendar usually don’t go back to a lunar calendar of twelve 30-day months, even though the older, simpler method would wipe the floor with the more accurate method (“What, so 1600 and 2000 are ‘leap years’ but 1700, 1800, and 1900 aren’t? VOTE NO! One month, one value!…” etc) if the latter had to pass a referendum before it could be trialled in practice.

      • I would add France, Zimbabwe, Liberia, Sierra Leone and the Ukraine to Tom’s list, and also note that referendums on transitioning to non-PR systems, which are presumably indicative of some support for change, have taken place in Poland, Romania, New Zealand and Italy. I suspect the relative novelty of PR in the Anglophone world reduces the coverage of campaigns for reform away from PR.

      • Agreed, Henry. Although not all of those reverted to FPTP (eg, France). Zimbabwe did, though – hey, No2AV, looking for a poster child? Found you one!

      • As it was, New Zealand’s FPTP supporters in the 1993 referendum cited Italy’s move away from PR a few months earlier in their campaign propaganda. However, they conveniently omitted the fact that the new Italian system retained a reduced proportional component, whose raison d’être was precisely to allow limited representation for smaller but significant parties unlikely to win plurality seats. I’m not certain that at the time Italian parliamentarians envisaged the emergence of political cartels as the main competitors under the new system, and that few of the smaller players would end up running on their own. In fact, as I’ve said before the most notable beneficiaries of the proportional component during the twelve years the system was in place were what remained of the once-mighty Christian Democrats in 1994.

  4. Why do polities insist on using majoritarian referenda to decide whether or not to abolish a majoritarian electoral system with proportional representation when they have enough support in Parliament to change the law? It’s not like they will get punished under the new electoral system.

  5. What if a Canadian providence just decided to change the electoral system to PR without a referendum to be used for the next election? Would the electoral system stick or be reversed when the opposition who is oppose to change abolish it if it wins the next election? Like what happen in France in 1986 or when BC used the AV for 1952 and 1953. Or will the system persist for a few elections, then the opposing party that is oppose to it wins and then abolishes or can’t because it can’t win a majority and it’s proposed coalition partner is opposed? Why did preferential voting persist in Australia despite at the time in 1918? Labor could have abolished it, but didn’t, but now today Labor benefits from this system as it gets preference flows from the Greens.

    Interesting article about BC use of AV; http://citeseerx.ist.psu.edu/viewdoc/download?doi=

    • I take an explicitly asymmetrical view on this. A referendum should be constitutionally mandatory to revert from a transferable to a less-transferable voting system or from an adequately proportional to an inadequately proportional one. But not vice versa. Legislatures should be morally and constitutionally free to reduce the percentage of wasted votes by ordinary legislation, without requiring a binding referendum.
      (Defining terms. 1. I mean “transferable” in the broadest sense, ie includes via runoffs or party lists as well as by voter-expressed preferential numbering. 2. Reducing the average district magnitude from 15 seats to 10 shouldn’t need a referendum, but reducing it from 3 to 1 should. 3. Single-seat AV and runoff are “transferable” — all parties below the quota can get their votes re-allocated, but since the quota is 50.01% this doesn’t help the smaller players. By contrast, PR List systems are proportional but not transferable – there are enough seats per district that every party with 4% to 8% will win representation, but those under the threshold [a quota, or one of the highest sub-quota averages or largest sub-quota remainders] get their votes binned.)
      The reason for this is not just that I support proportional and/or transferable voting methods over non-proportional and vote-splitting alternatives in general. It’s also specifically because the existing use of proportional and/or transferable voting systems greatly reduce the need for a referendum as a costly separate instrument for ascertaining the majority will, separate from ordinary legislative elections. In other words, if FPTP supporters ever do manage to convince a majority of the voters in Israel or Germany that FPTP is the way to go, then they don’t need a referendum to elicit the majority will. They simply use the system to give the pro-FPTP party or coalition a majority of seats. That party or coalition will need to have 50.01% of the votes or something very close to it.
      By contrast, if FPTP is the status quo, then it is entirely possible that a plurality party (which will almost certainly support continuing FPTP) well under 50% will win majorities of seats even if two or more sizeable opposition parties support a transferable and/or proportional system. The fact that “a majority of MPs elected under FPTP support FPTP” means almost nothing. Those MPs may represent (as Britain has shown in the last 50-60 years) only 40% or even 35% of those who cast ballots. A referendum is therefore necessary, otherwise issue may not be decided by anywhere near 50% of those who turn out to vote.

    • I’d think an electoral system change approved by referendum makes it less likely a party opposed to said change would seek to undo it upon winning office by means of a simple law, because that would be perceived and decried as “going against the will of the people.” If the voting electorate has previously approved the measure, it becomes almost de rigeur to hold a new referendum in order to repeal it. That said, if the party opposed to the new electoral system has an authoritarian disposition, its leaders may not feel bound to abide by such considerations upon attaining power.

    • Quebec’s governing party is committed to introducing PR without a referendum. We’ll see how that goes, politically–both whether they follow through (having won a majority of seats under FPTP) and, if they do, whether some future government feels just as entitled to reintroduce the old system (or something else).

      • “whether some future government feels just as entitled to reintroduce the old system” Surely that would require a party to win a majority of seats first – not impossible but very unlikely. If anything, Quebec politics are likely to fragment further under PR.

  6. There seems to be some kind of Law that says no matter how small a jurisdiction appears from the outside, it looks large to those inside it. Even in PEI, politicians who favour FPTP can oppose PR because districts will be “too big”.

    In debate on a motion to replace FPTP by MMP (PEI Fall 2016 Hansard, page 1640) we find this: “I think it’s very important that in our elected House that people have the opportunity to actually go door to door. It’s a challenge for me to make it to every house in my district. I still only have 3,400, 3,500 voters, but they’re usually a few minutes apart every house, so it takes a little bit more time.”

    As astronomers would put it, the election of Western Australia senators by STV probably is outside the event horizon of Islanders.

    • PEI is so small that the entire legislature could be elected by one multi member district using an open party list system, yet English speaking jurisdictions have an obsessive fetish with geographic representation, Single Member Districts boundaries are not naturally, they are artificial and do not always represent communities of interest.

      • PEI is so small that Western Australia has at least a dozen state and federal seats larger than PEI.

        When you’re ingrained to do things one way, you’re going to do it one way.

        Not that that makes your point invalid

  7. For PEI, if there is concern about the SMD being too large, then it would be better to increase the size of the assembly, but then that might defeat the electoral reform. Do voters care about the size of an assembly and the number of politicians to represent them or do they want small assembly sizes because they hate politicians and wish there were fewer of them, but one can’t have no assembly or an assembly of one then it wouldn’t be a democracy. There has to be a minimum number for an assembly size, most city councils have a minimum 3-5. Maybe we would be better off not having elections or have one chamber use sortition and assembly by jury, and let every citizen vote for every law on smart phones.

    • Don’t forget the single member county councils in Georgia.

      I agree that it seems that people do not want more politicians. The feeling seems to be that if there are more elected officials, those officials will want to more things, and will need expensive staffs and “research trips” to prepare for whatever programs they will create.

      I may be in the minority, but I feel it should be opposite. An elected body should have more people on, including the executive if elected separately. More voters means it will be harder to whip up votes for things that are unpopular, harder to slip things through without anyone noticing, and easier for a backbench rebellion to slow things down.

      • Yes. And, in addition, I think smaller SMD’s are on average less disproportional than larger ones. At least they give smaller parties a better chance. But you’re only in a minority in the larger population. I suspect that almost all readers of this blog agree with you.

      • I don’t know about the single-member country councils in Georgia. Do tell, Mark!

        Bob, yes, geographically smaller single-seat districts, for a given territorial jurisdiction, would certainly tend to be more proportional, on average. The logic is that means a bigger assembly, hence a bigger seat product. A bigger seat product means more parties expected to win, and reduced deviation from proportionality.

        What advocates of assembly size increases, myself included, sometimes forget to mention is that this is bought at a cost of fewer competitive seats (again, on average). That is, the newer, smaller districts are likely to be more homogenous. This is pretty much necessary to generate lower disproportionality: not only do the added districts mean more opportunities for smaller parties, but they also simply mean fewer votes cast for losers as winning parties tend to win by wider margins. Brian Frederick’s book on the size of the US House goes into these matters at some length.

    • In the specific case of PEI, the legislature has only has 27 members since 1996. Before that it was 32, even though the population has presumably increased or at least remained stable in the meantime. Therefore few people could have any complaints about a modest increase in the size of the legislature for the sake of balancing local representation with proportionality.

      • I think that might be a hard sell, given that PEI’s legislature already has the lowest ratio of population to members of the legislature by a long way (PEI has about 5300 people per MLA, with the next lowest being Newfoundland with 13000).

      • Large legislators are good. Expanding that when people are paying $20 a head per year for a representative may be a hard sell.

      • By cube root law, PEI should double the size of its assembly.

        Even if one believes the cube root does not apply at sub-national level (plausible, but I’d want a quantitive logic behind that before agreeing), it surely should be increased by at least several seats.

      • The ACT legislative assembly increased from 17 to 25 members for a population of 406,692 in 2016 and that was within the rather tight constraint of seeking an odd number of districts each returning an odd number of members that is at least 5. There was some, but not very much, controversy. The ACT area is 2,358 km² against 5,660 km² for PEI. The ACT has no municipal governments.

      • The ACT case Alan mentions is interesting. In percentage terms, going from 17 to 25 is quite a large one-step increase!

        However, by cube root, the current size is still very small. It should be 74. I assume the absence of municipal governments makes the theoretical case for larger size stronger, rather than weaker.

      • The other issue is the ministry, in parliamentary systems that require ministers to be assembly members. The ACT has 8 ministers, which is quite a lot to face an assembly of 17, especially an assembly where the speaker only votes if there is a tie.

  8. The opposition to increasing existing assembly sizes — even when warranted based on the cube root rule — clearly involves distrust of politicians in general among the population at large, as already noted in this thread. Does anyone think it also involves some level of recognition by elites that larger assemblies are somewhat less disporportional? What other motivations might elites and professional officer holders have for keeping assembly sizes artificially low?

    • Smaller assemblies are much easier to whip and much easier to a advocate. The small assembly fits into a popular misconception that assembly size has something to do with total public sector size. Broadly, popular misconceptions prevail when they are so beneficial to political elites, and when said elites command an almost infinite number of special majorities that they can impose unilaterally to entrench tot he existing system.

      Let us imagine in the US the XIV Amendment had been put to a popular vote where it required majorities in 60% of all congressional districts or the Voting Rights Act had been put to a popular vote where it required 60% to pass.

    • Bob, I think the reasons you and Alan give are probably correct. (I wish I knew for sure.)

      Of course, any explanation of undersized assemblies should also account for those cases where the assembly is oversized. The UK obviously comes to mind, but several other European countries are also on the high side. Then there is New Hampshire. Hungary’s was also quite oversized, but in the recent package of constitutional reforms, at least the assembly size was one thing they got right. (I mean, right in a narrow cube-root sense; it was one of several changes that enhanced disproportionality and thereby helped enhance Orban’s power.)

      • I suppose many readers know that one of the slogans the pro-FPTP side tried in New Zealand in the 1990s was “MMP means more MPs”.

        Any campaign against assembly-size increase surely starts off with a substantial advantage because such slogans are a lot likelier to resonate than any that one might cook up to defend an increase (or oppose a decrease).

      • Alan’s point that smaller bodies are easier to whip and easier to lobby strikes me as right on target. I hadn’t thought about it that way. Thanks!

  9. Manuel is correct. There’s a strong status quo bias. Two consecutive democracy-failures, where the existing system blatantly puts the wrong person in power, is what allows election reform to be passed. Even one failure is enough in some more progressive places like Fargo ND, but most places seem to require two.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.