Quebec to have electoral reform referendum

Per CTV News Montreal, the CAQ governing party in Quebec promises a referendum on a specific electoral-reform proposal to be held concurrent with the next general election in 2022.

The CAQ government on Wednesday introduced its electoral reform law, but backed away from its 2018 campaign promise to have it in place in time for the next general election.

The system is a form of mixed-member proportional representation (MMP), but a complex one. And not very proportional.

The number of seats in the National Assembly would remain 125. The new electoral law would divide Quebec into 80 larger electoral divisions that will mainly be the same as federal ridings, and 17 wider regions:

  • 80 MNAs would be elected according to the current system

  • 45 MNAs on a list of candidates would be divided among the parties, according to the percentage of votes obtained in the 17 regions

Thirty six percent of seats for the list tier is certainly on the small side, and the proportionality would be reduced further by that list tier itself being districted. Note that the mean number of seats per compensation region is only 2.65.

Based on further detail that I learned from Manuel on Twitter (and that I trust he will not mind my sharing here), we can see yet more ways that this proposal is designed to limit proportionality.

…it’s a very constrained implementation of PR, limited by a provincial-level 10% (yes, ten percent) threshold; and districted MMP in seventeen regions, with a new variation of the D’Hondt rule that skews seat distributions in favor of the larger parties.

Regarding the seat allocation method for the compensation seats (which will limit how compensatory it actually will be):

In Scotland and Wales the modified D’Hondt divisors are N+1,N+2,N+3, and so on, where N = number of single-member seats. In the Quebec proposal, they are N/2 + 1, N/2 + 2, N/2 + 3, N/2 being *half* the number of single-member seats, rounded up, and resulting in lower divisors.

One additional detail: the bill provides for separate allocations of single-member and PR list seats among regions – according to the number of registered voters – which guarantee all but one region a minimum of two seats. This would cost Montreal – a PLQ bastion – three seats.

I would still consider this MMP*, as there is a compensation mechanism. I am on record as considering even the Jenkins Commission proposal in the UK to be MMP, albeit with lots of caveats given it was also designed to be about as weak on the P as could be.

Regarding the election of 2018 when the CAQ came into power–surprisingly, with an absolute majority of seats–and the electoral-reform promises made at the time, see this earlier planting.

  • * UPDATE: I am now not sure about this; I need more time to think it through, and that will have to wait till some time in 5780! In the meantime, see this Twitter thread. I think the issue hinges on whether the “modification” to D’Hondt is actually more like Imperiali. While Taagepera and I list Imperiali divisors in our book as being part of the family of PR allocation formulas, we both now believe that it should not be. We were prompted to this view by an email exchange earlier in 2019 with Steven Verbanck (regular F&V commenter). Anyway, to be continued…

38 thoughts on “Quebec to have electoral reform referendum

  1. Pingback: Rediscovering an old publication | Fruits and Votes

  2. It’s been noted elsewhere that the notional result of the 2018 provincial election under the new system (apparently using federal electoral district boundaries) would have been (with change relative to the actual outcome in parentheses) CAQ 60 (-14), PLQ 36 (+5), PQ 15 (+5) and QS 14 (+4). CAQ would have lost its overall majority, but all the same would have won 48% of the National Assembly’s 125 seats with 37.4% of the vote.

    At the regional level, the result in the Capitale-Nationale shows how the proposed system works to benefit CAQ. The region would have eleven seats (same as now): seven single-member districts (coinciding with the number of federal election districts) and four PR list seats. In 2018 the four qualifying parties had the following results there:

    CAQ – 174,501 (43.4%)
    PLQ – 88,514 (22.0%)
    PQ – 47,103 (11.7%)
    QS – 67,921 (16.9%)

    The transposition of votes exercise I carried out earlier this year shows CAQ would have won six division i.e. single-member seats in the region to one for QS and none for PLQ or PQ; the actual election outcome was 8 CAQ, 1 PLQ, 0 PQ and 2 QS.

    Under the Scottish/Welsh modified D’Hondt rule, the top four quotients would have been:

    88,514 PLQ = 88,514 / (0 + 1)
    47,103 PQ = 47,103 / (0 + 1)
    44,257 PLQ = 88,514 / (1 + 1)
    33,960 QS = 67,921 / (1 + 1)

    As such, CAQ would have had 6 seats, PLQ 2, PQ 1 and QS 2, with CAQ securing 54.5% of the seats with 43.4% of the vote.

    However, under the Quebec proposal modified D’Hondt rule the top four quotients would be:

    88,514 PLQ = 88,514 / (0 + 1)
    47,103 PQ = 47,103 / (0 + 1)
    44,257 PLQ = 88,514 / (1 + 1)
    43,625 CAQ = 174,501 / (3 + 1)

    Thus CAQ would have won 7 seats, PLQ 2, PQ 1 and QS 1. Even though CAQ was already over-represented with six seats, it would have secured seven (63.6%) as the lower divisor (3 + 1 = 4 instead of 6 + 1 = 7) would have allowed the party to have the fourth largest quotient, losing just one seat instead of two with respect to the actual outcome.

    Moreover, the Capitale-Nationale would be just one of three regions with more than ten seats in all (the other two being Montréal with 24 and Montérégie with 22); the rest of the regions would have nine or fewer seats, and just one to three PR list seats (none in the case of the vast, sparsely populated Nord-du-Québec region), so the corresponding seat distributions would in all likelihood be even less proportionate.

    Premier Legault has stated that reform of the electoral system has proved to be more complex than anticipated. Perhaps so, but I might add that evidently letting go of an unexpected and comfortable parliamentary majority has also proved to be more complex than anticipated, as shown by the proposed reform itself.


    • I should caution that in several regions the single-member divisions under the proposed MMP electoral system may be significantly different from the existing federal electoral districts, contrary to press reports. Beyond the exceptions provided for Ungava and the Magdalen Islands, the allocation of single-member seats among regions would appear to preclude the existence of seats crossing regional boundaries, a number of which exist currently at both federal and provincial levels. Moreover, in some regions the assigned number of single-member seats does not equal its corresponding number of federal electoral districts. This is the case in Montreal, which would be allocated 16 single-member divisions, or two fewer than its 18 federal election districts.

      The transposition of votes I carried out for last year’s National Assembly election had PLQ prevailing 13 federal districts in Montreal, to 4 for QS and 1 for CAQ; PLQ won 42.9% of the vote in the region, to 21.9% for QS, 17% for CAQ and 12% for PQ. Meanwhile, reducing the number of Montreal divisions from 18 to 16 showed two groups of four contiguous federal districts each with significantly below average electorate totals, thus making them the most likely targets for consolidation into six divisions. All but one of these districts had notional PLQ majorities, usually very large, so the two eliminated seats would likely come at the expense of that party. The remaining district has a notional QS majority, but it doesn’t have a particularly large electorate deficit, and could be expanded along existing provincial (i.e. National Assembly) boundaries; in that case, QS would retain a narrow majority in that district. In all, the notional distribution of National Assembly SMD seats in Montreal under the proposed system would be 11 PLQ, 4 QS and 1 CAQ.

      Consequently, the distribution of all 24 seats in Montreal (currently 27) under various electoral systems would be as follows:

      FPTP (actual outcome): 2 CAQ / 19 PLQ / 0 PQ / 6 QS
      Scottish/Welsh MMP: 4 CAQ / 11 PLQ / 3 PQ / 6 QS
      Quebec MMP: 3 CAQ / 13 PLQ / 2 PQ / 6 QS
      MMM (parallel): 2 CAQ / 15 PLQ / 1 PQ / 6 QS

      Note that the modified D’Hondt rule would work to the benefit of PLQ in Montreal, which would pick up two additional list seats; one of these would come from PQ and the other at the expense of CAQ, but for Quebec’s ruling party the impact would be minimal and under the proposed system it would pick up an extra seat (relative to pure FPTP) in one of its weakest regions in the province. Meanwhile, QS would win the same amount of seats under all systems. Note as well that in the specific case of Montreal, the modified D’Hondt rule produces a distribution of seats that’s exactly the average of the figures obtained under Scottish/Welsh MMP and MMM. However, I should note that simulations I have run of the system for elections elsewhere show that’s not always the case.

      In conclusion, the proposed modified D’Hondt rule would have a distinct impact even in Quebec’s largest region seat-wise, even though it wouldn’t favor the ruling party in that part of the province.


      • And to add fuel to the debate about whether Quebec’s proposed electoral reform is truly an implementation of MMP, I re-ran the results of last October’s National Assembly election under Hungary’s current electoral system, with 78 single-member constituencies – the existing federal election districts – and 47 party list seats. Although the Hungarian system has a nationwide compensatory mechanism, it’s very weak and the system is not considered MMP. As it turns out, the distribution of National Assembly seats under the Hungarian system, compared to the published seat estimates under the proposed system would be CAQ 63 (+3), PLQ 33 (-3), PQ 15 (=) and QS 14 (=). Reach your own conclusions.


    • Although I think it’s worth noting that the single-member districts and PR districts with a 10% threshold are likely to work together in such a way that most political parties won’t miss out on representation in one form or another. A party with less than 10% support which draws its support from a particular geographic area, like the anglophone Equality Party (which won 3.7% and 4 SMD seats in 1989) or the Kurdish nationalists in Turkey, will be able to win SMD seats. The parties that will lose out entirely on representation will be those with support levels between 5-10% and who have their support spread relatively evenly across Quebec.


  3. “N+1,N+2,N+3, and so on, where N = number of single-member seats”

    How is that, in and of itself, modified D’Hondt?


    • In standard D’Hondt the divisors are 1,2,3, and so on. However, in Scotland and Wales a party may win more SMC seats in a given electoral region than the number of list seats it would be entitled to under regular D’Hondt. For example, in the 1999 and 2003 Scottish Parliament elections, Labour won all ten constituencies in the Glasgow region, even though it received 43.9% of the party list vote in 1999 and 37.9% in 2003. Since Scottish/Welsh MMP does not provide for overhang seats – that is, neither devolved legislature can be expanded to accommodate a party winning more constituency seats than it would be proportionally entitled to – and a regular D’Hondt distribution of the region’s seventeen seats (ten SMC and seven party list) would have resulted in Labour being allocated fewer seats than its constituency seat total, the D’Hondt rule was modified to make the distribution in each region as proportional as possible within the constraints of the SMC outcome, while allowing parties to keep all their SMC seats. In the case of Glasgow in 1999 and 2003, this meant the D’Hondt divisors for Labour were 10+1, 10+2, 10+3, and so on, and 1,2,3 for all the other parties, since none of them won any constituency seats in the region. The higher divisors lowered Labour’s D’Hondt quotients, and the party won no list seats in the region, which made sense since by virtue of its constituency victories it already held 58.8% of the region’s seats (10 of 17).


      • You can actually characterize the system compared to D’Hondt by the coefficient k they are applying to N. In traditional Scottish MMP the coefficient is 1 – full compensation. In MMM (ie Italy) the coefficient is 0 – no compensation. This is literally the arithmetic mean of the two.


      • That’s interesting.

        Just one amendment to that: The Italian system actually had a complex partial compensation mechanism, involving adjusting list votes according to nominal contest outcomes. Hungary has something similar (although the precise mechanism is different).

        Pure “parallel” allocation, and hence the most basic MMM, would be on the Japanese model. The list votes are not adjusted in any way, nor as the list seats per party in any way dependent on nominal (single-seat district) outcomes. I don’t think this changes your point, Ryan. I am just clarifying the parallels (so to speak!) to other systems.

        Also, while I can’t claim my mind is yet made up about the characterization of the Quebec proposal, I just added an update to the original post with a disclaimer regarding my calling it MMP.


      • Thanks for the clarification! Does the current Italian parallel system still have the partial compensation though? I was under the impression they had switched to a more “traditional” MMM system for their most recent election.

        I’m not sure what to call this system to be frank. MMP may be the best option for selling it though I guess.


      • The French name may be more straight forward. In French MMP is le mode de scrutin mixte compensatoire. So maybe this is le mode de scrutin mixte partiallement compensatoire? Demi-compensatoire?


        • The only (tenuous) link between the FPTP and PR components in Italy’s current MMM system is that votes cast for a coalition FPTP candidate rather than for one of the parties backing said candidate are proportionally distributed at the single-member college level among coalition partners. However, less than 4% of valid ballots were cast in that manner in the general election last year.


      • The use of small regions makes overhangs more likely. The adjustment for overhangs takes from list parties to give to the SMP dominant party. Granting overhangs by taking them from the list has about double the anti=proportionality effect as just allowing the extra seats. And the small regions accidentally (on purpose) make those very overhanges more likely. In 2021 SNP won two list seats, AND diverted other list seats to it to fit its overhangs.


  4. I think it would be correct to state that the 17 regions of Québec already exist in law, and are not being created simply for the purposes of changing the electoral system.


  5. Since so many people have posted projections, I will venture in. My assumptions on the numbers of ridings per region are based on a Quebec post I have not verified. There seems to be some dispute on how to deal with Bertrand and Côte-du-Sud, and of course every one has their assumptions on who would win the larger local ridings. However, the overall picture seems to be about this:
    CAQ: FPTP 74, perfect Quebec-wide PR 49, highest remainder with these districts and regions 54, this model 60.
    Liberals: FPTP 31, perfect Quebec-wide PR 33, highest remainder with these districts and regions 31, this model 31.
    PQ: FPTP 10, perfect Quebec-wide PR 22, highest remainder with these districts and regions 21, this model 19.
    QS: FPTP 10, perfect Quebec-wide PR 21, highest remainder with these districts and regions 19, this model 15.

    So the moderate (semi-proportional) calculation system (whatever one calls it) is roughly half of the problem, while the small regions and the 36% compensatory seats is roughly the other half. The 36% had been agreed on. The small regions will be disputed but may be popular in those regions, and they at least give QS seats in six regions where they are now unrepresented. The calculation system is the price of CAQ support. Will QS and the PQ decide this is worth a campaign for a YES vote, it would have given the CAQ less than half the seats, get moderate PR and improve it later?


    • To correct the record: I no longer project the numbers I first did. But my corrected projection does not match the government’s. So I will shut up until I can sort it out.

      A few things are clear:
      1. The government has indeed tabled a bill to create a proportional voting system, as promised and on time.
      2. The referendum question can be answered with a yes, invites agreement, and sounds to a voter like the government must know what it’s doing: “Do you agree with replacing the first-past-the-post electoral system by the mixed electoral system with regional compensation set out in the Act to establish a new electoral system? Yes/No.” And no super-majority is required.
      3. Some reformers are looking to improve the model during the public hearings:
      • 17 regions, four so small they have only one regional compensatory seat, and one is an exceptional single-member district.
      • Quebec-wide threshold of 10%
      • Dual candidacy prohibited
      • Calculation method providing a winner’s bonus
      • Distortions reduced by only 40%
      • Majority governments are possible with less than 50% of the votes
      • No binding measure for women’s representation.
      • No measure to represent diversity.
      • The referendum is during an election campaign.
      • It will not apply for the 2022 election.


      • I’ve done simulations of the 2018 election outcome for some regions, but not for all of Quebec. Some were easy because the number of FPTP seats matched the corresponding number of federal election districts, and the latter didn’t cross regional boundaries. Others are trickier, but in the end turned out to be far easier than initially anticipated. For example, the Montérégie region would have 14 single-member divisions, but contains 11 whole FEDs and parts of another four. Those parts haven’t enough registered voters to carve three whole divisions out of them – even factoring the 25% variation allowance in Quebec’s electoral law – so some of the adjoining divisions would have to be modified as well. However, PLQ only has notional majorities in two of the 11 whole FEDs in the region, which CAQ carried by a very large margin, and the areas where boundary reviews would have to be undertaken went solidly for CAQ. That said, the boundary review could have a domino effect which in turn could potentially affect the Vaudreuil–Soulanges federal election district, which has a notional PLQ majority of 4,680. I’m strongly inclined to believe that majority would hold, but can’t say I’m 100% certain yet, so the SMD seat distribution in the region could be 12 or 13 CAQ and 1 or 2 PLQ.

        Alas, in a curious turn of events CAQ would win 13 of 22 regional seats all the same: if it won twelve SMD seats it would be entitled to one of eight regional list seats, but not if it won in thirteen divisions, even with the modified D’Hondt rule in place. All the same, CAQ would be over-represented in Quebec’s second largest region population-wise, securing 59.1% of the seats with 41.8% of the vote.

        As for improving the proposed system, I’ll be blunt: CAQ and Premier Legault have zero incentive to change anything so long as polls continue to indicate they will retain a comfortable National Assembly majority under the current FPTP system. However, if CAQ’s fortunes were to take a sharp turn for the worse in 2020 and 2021, and the ruling party appeared likely to be defeated in 2022 – particularly by a large margin – what we’ve witnessed so far leaves no doubt in my mind that they would rush to enact most if not all of the suggested improvements, and even scrap the referendum if push comes to shove.


      • I have now carried out two region-by-region simulations of Quebec’s 2018 National Assembly election under the proposed mixed member compensatory electoral system: one which apportions the SMC and PR seats among regions by population, and the other by number of registered voters, the latter as set forth in the province’s electoral law.

        Although I initially obtained a slightly different overall distribution of seats by party under the population method, I was able to replicate the figures published by the government of Quebec after reassigning one SMC to a different party in three regions. While these changes were at variance with my estimates, all of them appear to be more or less plausible.

        However, this brings me to a far more serious issue that came up in the process, namely that the simulation of the 2018 provincial election based on the number of registered voters method gave CAQ 64 of 125 seats, for an absolute majority of three. Of course, it’s entirely possible Premier Legault’s government is not aware of this, but it’s just as possible that not only they are, but that they deliberately used the population method – not contemplated by Quebec’s electoral law – for one simple reason: presenting a simulation that showed they had come up with a supposedly proportional system that nonetheless allowed them to keep an absolute majority of 51.2% of the seats with 37.4% of the vote would not have gone down well with the opposition parties and public opinion, so they had to come up with more presentable numbers for public consumption. If true – and that’s a very big if – it speaks volumes about Premier Legault and his government.

        Now, I want to make it very clear I’m still in the process of re-checking my data, but so far I haven’t found anything that indicates I made a mistake. Also, applying the adjustment of one SMC seat in three regions under the registered voters method still leaves CAQ with 63 of 125 seats, for an absolute majority of one – and the same number of seats it would have obtained under the current Hungarian electoral system.


      • @Manuel Alvarez-Rivera: with the seat distribution of 2018 seeded from numbers of past communication from Élections Québec, and votes amongst overlapping territories supposed uniformly distributed along their respective regions and finally imitating 2018 distortions for the local seats, I’ve also computed similar results:
        pq: 15 qs: 13 caq: 64 plq: 33

        Me and wilfredday have compared our computation without finding mistakes in the proceeding and I also have produced a second independent code to check my own numbers. Thus, at this point I put some trust on these numbers. Though, Elections Quebec might have more precise results for local and regional vote, by using geographical data of each of the polling stations, say.

        My next steps are going to compute a few voting case possibilities to better understand and commensurate the behaviors of this semi-parallel/semi-proportional voting system.


        • @user8019

          Thanks for letting me know. I find Elections Quebec’s reluctance to release full details about their published simulation very suspicious. At any rate, the results of my simulation are almost identical to yours: CAQ 64, PLQ 32, PQ 16 and QS 13. That said, my methodology was more elaborate. As I’ve said before, earlier this year I transposed the results of last October’s National Assembly election in Quebec to the province’s 78 federal election districts, and that data proved quite useful in many cases where the federal districts are contained within regional boundaries, even if other districts covering a given region are not. For example, in the case of the Estrie region, which would have three single-member division seats under the new system, the transposition of votes shows QS would have a notional majority of 1,269 in the federal district of Sherbrooke. However, CAQ prevailed in the rest of that region by a very large margin, so the other two divisions there would have almost certainly had notional CAQ majorities, irrespective of where the line was drawn.

          In addition, I used polling station-level data to determine regional results in the half-dozen cases where existing provincial constituencies crossed regional boundaries, although I still had to apportion special votes and in some cases advance votes in advance polling stations covering regular polling stations in more than one region.


        • Here’s the notional distribution of National Assembly seats under the proposed mixed-member compensatory system, with seats distributed among regions by their total number of registered voters. Apologies if the table doesn’t display correctly.







          Note that even though CAQ would have won 62.5% of single-member seats with 37.4% of the vote, the proposed system would have allowed the party to receive 31.1% of the regional list seats.


      • After your comment, I took the time to gather the missing information from the map data & polling stations of Élections Québec and got the same results as yours for the overall tally, but with an alternative smd pick:
        smd | pq: 5 | qs: 5| caq: 51| plq: 19| autres: 0|
        list | pq: 11 | qs: 8| caq: 13| plq: 13| autres: 0|
        régional | pq: 16 | qs: 13| caq: 64| plq: 32| autres: 0|


        • Thanks again for replying. I was able to replicate your SMD and party list distributions with two changes in my SMD distribution: one in Lanaudière (from 4 CAQ and 1 PQ to 5 CAQ and 0 PQ) and another in Capitale-Nationale (from 1 QS to 1 PLQ, with 6 CAQ unchanged). These changes don’t alter the overall allocation of seats because PQ would pick up an extra list seat at the expense of CAQ in Lanaudière, while QS would gain a list seat from PLQ in Capitale-Nationale.

          The flipped SMD seat in Lanaudière would be the federal electoral district of Joliette, which has a notional PQ majority of just 38 votes; however, that majority is contingent upon the apportionment of special votes, where PQ had a majority of 358, which overturned the CAQ majority of 133 in regular polls (CAQ won a further majority of 133 in undivided advance polls, and of 54 votes in split advance polls). Given the narrow outcome and its dependence on a mathematical computation, the district could go either way, and the notional result should be taken with caution.

          Meanwhile, in the Capitale-Nationale the notional results in the federal electoral districts show QS had a notional majority of 3,225 in the federal electoral district of Québec; CAQ had notional majorities in the region’s remaining six districts, although PLQ came within 2,612 votes of a notional majority in the federal district of Louis-Hébert. That said, a majority PLQ district in the region is possible with a different configuration of provincial divisions under the proposed system.

          At any rate, I have identified 43 federal districts (of 78) which could be used as provincial divisions under the proposed mixed-member compensatory system, while nine federal districts cross regional boundaries and are not suitable for use under the new system. Of the remaining 26 federal districts, most could also be used with boundary changes as provincial districts, but some would have to be broken up. In fact, one could say the assertion that the federal election districts will serve as the basis of the new provincial divisions under the proposed system is at best a half-truth.

          In all, besides Joliette I have notional PQ majorities in the federal districts of Jonquière (in the region of Saguenay–Lac-Saint-Jean) and Manicouagan (covering the entire Côte-Nord region); in the Bas-Saint-Laurent district resulting from the merger of Matane-Matapédia and Rimouski (or at least part of the latter); and in the two single-member divisions of the Gaspésie–Îles-de-la-Madeleine region, one being the existing provincial division of Îles-de-la-Madeleine, and the other by necessity the merger of the existing Gaspé and Bonaventure provincial divisions (a merger which probably won’t go down well in the peninsula, I might add).

          Meanwhile, besides the Capitale-Nationale federal district of Québec, I have notional QS majorities in the Montréal federal election districts of
          Hochelaga, Laurier–Sainte-Marie, Papineau and Rosemont–La Petite-Patrie; and in the federal district of Sherbrooke in Estrie. The reconfiguration of the 18 Montréal federal districts into sixteen provincial districts could affect eleven of the region’s existing federal districts, among them Laurier–Sainte-Marie and Papineau. However, QS has a very large notional majority in the former, unlikely to be affected by comparatively minor changes in the district’s boundaries (in fact, the QS majority could actually increase after the boundary review), while a reconfiguration exercise in which Papineau was expanded to cover the existing provincial districts of Laurier-Dorion and Viau still left QS with a narrow notional majority of 464 (the current federal district of Papineau covers all of Viau and part of Laurier-Dorion).

          By the way, in a previous comment I had erroneously stated that the region of Montérégie had eleven whole federal districts and parts of another four, which didn’t have enough registered voters to form three new provincial divisions. In fact, it has twelve whole districts, but one would have to be broken up to keep the remaining eleven intact, and that district along with the parts of three other federal districts in the eastern side of the region have more than enough registered voters for three new divisions – all of them with large notional CAQ majorities, irrespective of where the boundary lines are drawn. Consequently, I’m now quite certain about the previously tentative distribution of 12 CAQ and 2 PLQ notional SMD majorities in the Montérégie region.


      • “At any rate, I have identified 43 federal districts (of 78) which could be used as provincial divisions under the proposed mixed-member compensatory system, while nine federal districts cross regional boundaries and are not suitable for use under the new system. Of the remaining 26 federal districts, most could also be used with boundary changes as provincial districts, but some would have to be broken up. In fact, one could say the assertion that the federal election districts will serve as the basis of the new provincial divisions under the proposed system is at best a half-truth.”

        The region boundaries correspond to the provincial administrative regions: (Québec’s gov offer shapefile of these boundaries somewhere on the web). As for the districts, the “Secrétariat à la réforme des institutions démocratiques” told me that they haven’t any complete map at the moment. The federal 78 districts was a previous suggestion of MDN, but as you notice, Élections Québec wouldn’t be able to use this map as is with this draft law.


        • Yes, my understanding is that the electoral regions would be the administrative regions. Elections Quebec shows the regional boundaries on the provincial division maps available on its website, and I have the boundary file of the regions from Statistics Canada (where they are listed as “economic regions”). The CTV article cited by MSS notes that “the new electoral law would divide Quebec into 80 larger electoral divisions that will mainly be the same as federal ridings,” so it would appear that suggestion was adopted in principle.

          I don’t have a complete map at the moment either for the proposed system, and truth be told in some regions (or parts thereof) it isn’t really necessary to have one right now to obtain accurate estimates. For example, in Centre-du-Québec CAQ swept with 57% of the vote last October and won every regular poll there, including those in the part of the Johnson division located within the region. With that kind of result it’s fairly obvious that no matter where the boundary lines are established, all three three single-member divisions in the region will have notional CAQ majorities.

          However, for the region of Outaouais, closely fought by CAQ and PLQ, I did come up with a complete map. In that case I was able to use two complete federal districts, namely Hull–Aylmer and Pontiac, both with notional PLQ majorities. However, I couldn’t use the entire federal district of Gatineau because even though it lies fully within the regional borders, the rest of the region wouldn’t have had enough registered voters for its assigned fourth SMD, and the fourth federal district covering the region – Argenteuil–La Petite-Nation – extends into the Laurentides region. As it was, an easy solution emerged: the existing provincial division of Papineau (not to be confused with the Montreal federal district of the same name) had just enough registered voters to be retained in its current configuration; this was also the case with the part of the federal district of Gatineau that doesn’t overlap the provincial district of Papineau, so with that I had the required four divisions; as expected, the latter two districts had notional CAQ majorities, for an overall SMD distribution of 2 CAQ and 2 PLQ.

          That said, I do have a full map for another Quebec MMP exercise I described in a comment on the Quebec 2018 thread on this site a year ago, which also provided for the regional allocation of list seats, but had fewer single-member divisions (65), obtained by merging groups of two or three existing provincial divisions (while retaining nine single existing divisions as exceptional cases); combined the administrative regions into eight electoral regions; and delivered a highly proportional distribution of seats (nearly identical to provincial-wide PR). I got the impression Wilf Day wasn’t too fond of it because it was districted MMP, but little did he know Monsieur Legault and his government were concocting something he’d be even less fond of. Oh well, as they say in French, c’est la vie. At any rate, a compromise measure could be developed from that other exercise, or so I would think.


        • My transposition of Quebec’s 2018 National Assembly election results to the province’s 78 federal electoral districts can now be accessed on Google Sheets here. Even though Quebec has different party systems at the federal and provincial levels, and the notional results of last year’s provincial election were primarily intended to assist in determining the impact of the proposed mixed-member compensatory system, they may also prove to be relevant in the upcoming federal election.


  6. Apparently the Quebec government handed out their own simulation last week, which I cannot find online, stating: if the proposed voting method was applied to the 2018 Quebec election results, Legault’s Coalition Avenir Quebec would have had 60 seats instead of 74 and formed a minority government, with the Liberals at 36 seats instead of 31, the Parti Quebecois at 15 seats instead of 10 and Quebec Solidaire with 14 seats instead of 10. Since I don’t know how they came to these numbers, I cannot dispute them.


    • I found the simulation furnished by the Quebec government – which I’ve cited on my previous comments – on an article in a French-language site, available here. The image displayed in the article doesn’t show the full document, but from a partially cropped explanation at the top it would seem they used the existing federal electoral districts. If so, it’s not clear how they dealt with federal districts crossing regional boundaries, or with regions such as Montreal which on the basis of 2018 electorate figures would have had fewer single-member National Assembly districts than federal electoral districts.


      • The difference in Montreal even today is that the federal boundaries use population figures, while Quebec uses voters on the voters list. Montreal must have a higher percentage of under-18s and non-citizens than the rest of the province. That’s why it has 23.1% of the federal ridings but only 21.6% of the Quebec ridings. However, it may well be that this model will shrink this to 19.2%, but I do not know if the Ministry has published this figure.


      • Said journalist from La Presse has been asked for the full document, but he replied that Élection Québec did not produced a more detail view of their simulation, that the cropped part was all they got. I’ve ask Election Quebec myself for more details but they did not call back yet.


  7. Is there any chance of this referendum passing? Is there no need for electoral reform in Canada and its providences? The Quebec parliament is quite fragmented for using a FPTP. Does Duverger’s Law apply here?

    NZ electoral reform referendum is a fluke as both major parties thought they could defeat it as with Brexit for the UK. The wrath of referendums produce unexpected results.


  8. Wednesday was the first of five days of legislative hearings on Bill 39, with many citizen groups and electoral reform groups making presentations. This bill was the result of a four-party agreement in 2018, when no one expected a majority government. The CAQ government is trying to maintain that spirit Here is part of the Minister’s opening statement (courtesy of google translate):
    “”These consultations will indeed give us the opportunity to hear interesting and diverse opinions on the reform of the voting system proposed by the government. They will undoubtedly allow us to improve this bill which now belongs to all of us. It is with openness, as I said, that I will welcome, that we will welcome the different opinions that we will hear over the next few days.
    “Many have tried to carry out this reform in recent years. Several governments have promised to do so. It’s been 40 years … more than 40 years that we have talked about it, and someone said to me rightly: If it had been simple, if it had been easy, because this is a very laudable principle, an improved democracy, we would have done it already. Our government has taken an important first step towards historic reform. From now on, it is together, collectively, that we must go the rest of the way to achieve this.”


  9. Pingback: Quebec electoral-reform referendum on hold | Fruits and Votes

Leave a Reply

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

You are commenting using your 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.