Canada confidence-and-supply agreement, and irresponsible opposition

The Canadian government of Prime Minister Justin Trudeau and his Liberal Party have forged a confidence-and-supply agreement with the New Democratic Party (NDP). Under the terms of the deal, set to run thorough June, 2025, the NDP publicly commits to supporting the Liberal minority cabinet on budget and confidence votes, in exchange for the government advancing some NDP policy priorities. Seven policy areas are mentioned in the agreement itself. One of the key priorities–in fact, the first item in the list–is a dental care program, which has been a campaign pledge of the NDP in recent elections. There are proposals for “making democracy work for people,” but in case anyone is wondering, no, electoral system reform is not on the list.

Predictably, figures within the opposition Conservative Party is decrying the “back room” deal (as if it were not public–it is on the official government website, after all–and as if bargaining processes themselves were ever productive when carried out in a “front room” of scrutiny). But two candidates vying for the leadership of the Conservative Party have gone well beyond normal criticism of such a deal or the policies it will lead to. Jean Charest accused the government of ignoring the results of the election and of embracing an “anti-democratic” ideology. Patrick Brown said “the will of Canadians has been subverted.” (There are several contenders for the leadership; the party has been with an interim leader since shortly after its 2021 election defeat. Among the contenders, Charest and Brown would actually be considered relatively moderate!)

These are irresponsible statements, and are playing on ignorance about how parliamentary democracy works. Quite contrary to Charest’s statement, the Liberal-NDP agreement is precisely how democracy should work. Canada has a parliamentary form of government (far superior to the presidential form, by the way). Governments must maintain the confidence of the majority of elected representatives. If no party has a majority on its own—something the Conservatives have managed to achieve just once since 1988–then inter-party agreements stabilize the government and facilitate passage of policies favored by parties representing a majority. 

As for “will of Canadians” most political scientists would caution that there is no such thing as a general will. What makes democracy work is accountability at the next election, and cooperation between elections. That is what this agreement is about.

I believe it is not the first such agreement in Canadian history, although it is the first of many minority governments in some time to have such an explicit agreement. Broadly, there are three options when an election in a parliamentary system does not result in one party attaining more than half the seats. (1) Two or more parties can form a coalition cabinet; (2) one party can govern alone with a public commitment from a support party (or parties) in parliament; or (3) one can govern alone and seek case-by-case support from various parliamentary parties on specific bills and on the annual budget. One could add other options, as well, such as act as if you are the majority and dare the opposition to combine and vote you out, or call an early election and try to win a majority. The latter is, of course, precisely what Trudeau attempted in 2021, and the result ended up being hardly any different from that of the 2019 election.

This agreement represents the second of those three main options. The NDP gets no cabinet seats, and thus it is not a coalition. The NDP commits specifically not to vote against the government on budget or no-confidence votes, while the Liberals agree to take up some NDP policies. Thus the Liberal Party does not have to worry about the NDP joining with other parties against it, nor do the Liberals have to attempt to please the Conservatives or Bloc Quebecois in order to gain support for legislation. Thus it is firmly in the category of public commitment between a minority government and a support party. (Like all coalition and confidence-supply agreements, it is not legally binding, and either side could elect to break it at any time.)

It is worth noting that the Seat Product Model expects no-majority situations to be a regular occurrence. Given the district magnitude (1) and assembly size (currently 338), we should expect the leading party to average around 48% of the seats. Over time, this is very close to what we have observed. Since 1997, the first year the House of Commons had over 300 seats, the mean seat share of the largest party has been 49%, and the median has been 47.3%. The Liberals currently hold 47.0%. Five of nine elections since that time have resulted in less than 50% of seats, while the others have returned majority governments. Over the entire period since 1949, eleven of twenty four elections have returned minority situations. So roughly half of elections result in no majority, which is about what we would expect from a seat product that predicts about half the seats, on average, for the largest party.

Canadian party elites and the public thus should have got used to the idea that a majority is not the natural outcome of an election. They should further get used to the idea that, as a result, parties might strike deals to enable minority government to be stable and successful at implementing policy. Yet the habits of majoritarianism die hard, especially when both the empirical record and the Seat Product Model show that majorities always are a likely outcome, even if not necessarily the most likely at any given election. The majoritarian habit is even harder for the Conservatives to kick, given that they currently have no viable partners, and if they form minority government, their best hopes are either case-by-case deals or provoking early elections and hoping vote splits among other parties and wedge issues allow you to get a majority (both of which were practices during Stephen Harper’s two minorities before winning a majority in 2011). If the consecutive elections with similar results in 2019 and 2021 have convinced at least some party elites that a more consensual style is needed, it would mark an advance for Canadian democracy. But not an advance the Conservatives are going to find it easy to reconcile themselves to. Hence their resort to claiming inter-party cooperation is an illegitimate and anti-democratic practice, when that could hardly be farther from the truth.

30 thoughts on “Canada confidence-and-supply agreement, and irresponsible opposition

  1. Is the observed standard deviation of the % of seats of the biggest party predicted well as sqrt(s1*(1-s1)/n) or does it tend to be systematically bigger? Does it matter if the predicted s1 is within 2 standard deviations of .5?


    • I am not sure if this answers the question, but the average election under a simple (single-tier, single-round) electoral system is 1.04 times that predicted from the Seat Product Model, with a standard deviation of 0.29.

      If you mean that average I reported from elections since 1997 in Canada (0.490), the standard deviation on that is 0.055.

      (I happened to still have the dataset open, as I often do!)


      • Does the standard deviation vary inversely with the square-root of the number of contested seats or nonlinearly with the predicted s1?

        Sqrt(.49.51/n)=.055 implies an n of 83, so apparently the standard deviation exceeds what would be predicted from a Binomial probability distribution. Looks like it’s about twice as big, which makes sense since 834 ~=338.

        I know that measures of voter enthusiasm often decline with the size of M, I’m wondering if it’s better related to (.5-s1)/StdDev(s1, Seats).


  2. So, what’s it gonna take for the CPC to embrace the need to use PR as part of Canada’s electoral system?
    I presume a switch to #AV led by LP and NDP would make it even worse for the current off-center incarnation of the CPC?


    • I can’t see them going there. As things currently stand, their only coalition partner would seem to be the PPC. So they are much better off–from their own perspective–counting on voters to tire occasionally of the Liberals and/or center-left votes to split, allowing the Conservatives win majority government.

      On the potential for AV in Canada, see this:

      Is AV just FPTP on steroids?


      • O I know well the antipathy of non-LPers to #AV…

        My point is #AV would further force the CP to move to true center of Canadian politics, which would have feedback effects on the LP regardless.
        Since LP put the kibosh on PR, It’d be better to get #AV than sticking with FPTP which enables off-center ideologues in the CP to rely on guile to resist changes.


      • I personally think one of the largest barriers to proportional representation in Canada is that polls show almost half of Canadians think we have it already. Which may explain why we are such an outlier for the Seats Product Model. The way civics are taught in Grade 11 actually reinforces this by giving a proportional distribution of seats to the classroom model parliaments.

        In terms of the Conservatives coming around on the issue, I’d note some prominent Conservatives are giving the idea a second look. See below:

        Michelle Rempel Garner is backing Patrick Brown. He is in an informal pact with Jean Charest, who himself tried and failed to implement a very limited version of MMP for Quebec when premier.

        So we reformers are in a tough position with respect to who to cheer for.


  3. One question I’ve always had about minority governments in Canada, and I don’t know if it’s within the scope of Professor Shugart’s usual interest, is whether “the largest party forms a government and can count, under most circumstances, on the support of the opposition” counts as a constitutional convention. The one opinion I’ve received from a Canadian expert was that it didn’t, but to me it seems to have some superficial similarity to other things that would be called conventions.


    • It’s discussed at some length by Anne Twomey in The Veiled Sceptre: Reserve Powers of Heads of State in Westminster Systems (2018), Chapter 3. Her answer is negative. Twomey’s work was cited with approval by both sides and the judges in the UK supreme court prorogation case in 2019.

      To my knowledge this alleged convention has been advanced only by those high constitutional scholars, Tony Abbot and Stephen Harper, who both headed the largest parliamentary party but not a majority parliamentary party, at the time. The governor-general of Australia specifically declined advice to that effect from Abbot in 2010 when she appointed Julia Gillard prime minister. Gillard headed the second largest parliamentary party but did have majority parliamentary party support.


  4. Would STV be a good electoral system for Canada? Canada could use small district magnitude of 3, 4, and 5 like Ireland.


    • Small regions are not proportional enough – commissions and Citizens Assemblies that have considered STV have suggested a range of three to seven. When Northern Ireland used six it worked. But at the federal level, only 51% of our MPs are elected from communities with 5 or more MPs. We have too much geography for STV, federally. (It would work well for Vancouver City Council.) This is why every commission and study east of the Rockies has recommended MMP, with open regional lists for the 40% of MPs elected for top-up seats. Regions might range from 8 to 14.


      • Quebec, the largest Canadian province, is 58.35% the area of Western Australia, but WA successfully elects its legislative council and its federal senators by STV. Ontario, the second largest Canadian province, is 58.13% the area of Queensland but Queensland functions successfully as a single district for electing federal senators. NSW, with an area of 809,952 km sq, functions successfully as a single district for the election of 21 MLCs and in 1918—1927 elected its legislative assembly by STV in multimember districts ranging from 3 to 5.

        Far from finding the NSW legislative council remote from my concerns, as a minor party voter submerged in a sea of Coalition voters I’m quite happy to contact a remote, inaccessible, faraway, disinterested, distant and impersonal MLC who shares my political views than dealing with my local MLA whose views formed sometime in the Early Jurassic. It was a bit difficult contacting my MLC when the trip to Sydney took 3 months by bullock cart, but luckily I’ve recently discovered this thing called the telephone. I may get all newfangled and even start using email some time. The more daring MLCs have even been known to venture out of the Big Smoke occasionally and come talk to us remote voters,

        The argument from Canadian geography is grossly overstated. It is certainly a factor to be considered, but it is not the decisive factor that it is often claimed to be. And in any case it could be overcome by establishing low magnitude STV districts in remote areas.


      • How about introducing fractional MPs? QMPs would require 1/4th as many votes and have 1/4th as many duties and in the general assembly their votes would weigh 1/4th as much. They’d be part-time and mostly telecommute. A lot of other stuff would need to get worked out, but a rural province having ridings w/ 4 QMPs would be better than having 1 MP, as was the case in Alberta when PR was being used for urban but not rural areas.


      • Alan, in Scotland’s last election, as a Green party voter in a sea of other voters, you would not have been submerged: in seven of the eight regions you would have elected a Green Regional MSP, as well as a local constituency MSP, and about six other regional MSPs from two or three other parties. The best of both worlds (except for closed lists). The Law Commission of Canada, inspired by Scotland’s system, recommended we have that in Canada too, except for open lists.


      • I’m sympathetic to Alan’s point of view here but I do think it’s a problem for PR that these concerns are being brought by him in particular. If PR could be implemented by winning the support of voters who were far from their district medians (as a Greens voter in rural NSW is) and as such placed next to no value at all on being able to contact a local MP, then I think it would be a very compelling point. But most rural voters aren’t Greens supporters. A Nationals voter in a rural area who places any value at all on having a specific ‘local MP’ is going to be much more concerned about a proposal to get rid of that local MP given that they lack Alan’s intense partisan disagreement, and, as I say, those National voters are much more numerous.


      • Henry

        I am forced to preface my reply with the tragic words ‘With the greatest possible respect’ I am not a member of a political party. I did not vote Green at the last state election. I have friends who have lived here all their lives and do vote National for economic reasons but nevertheless take their environmental concerns to an MLC.

        There is no factual proposition in your comment that is in any way accurate.


      • Wilf

        Contrary to rumour I am neither a rusted on Greens voter nor unaware of (1) the existence of regional MSPs in Scotland nor (2) the somewhat metaphysical nature of the proposition that thou shalt not have MPs elected by region except when you do.


      • I didn’t know if you were always a Greens voter or a Greens member, didn’t say you were either a ‘rusted-on Greens voter’ or a Greens member, and I don’t think the substance of my point has much bearing on whether you were or not. But I don’t know how else to interpret the phrase “as a minor party voter submerged in a sea of Coalition voters” except as suggesting that your voting behaviour is generally atypical for your geographic area.

        My point is that rural minor party voters who prefer having a seven-seat electoral district (even if it is geographically large) because one of the representatives for that seven-seat district is an ideological ally are unlikely to be a very significant part of an electoral reform coalition. Of course this is contingent upon a bunch of things, such as the desire to have a specific local MP being real and significant, but I’m just trying to explain why it seems implausible that the factor you describe is going to be very significant in electoral reform campaigning/


      • Henry

        Had you intended not to assume ‘intense partisan disagreement* it may have been better not to write about my ‘intense partisan disagreement’. Equally, there is no reason to assume Far West NSW would be a district of magnitude 7. We got on quite successfully as an STV district of magnitude 3 in the 1920s. All three MLAs for the area voted against abolishing STVPR.


      • Henry / Alan – I’m pretty sure the median distance to your closest MP would actually go down with multimember districts. The ability to cross previous boundaries means a large set of voters can go to a closer MP that previously would have been in a neighbouring district. That’s actually a really significant effect in some of Canada’s more rural districts too. For example, in Desnethé—Missinippi—Churchill River, if you’re in the eastern part of the district, you can drive to the neighbouring district MP’s office in about 3 hours. The trip to your own MPs’ office is about 8 hours each way. And you quite literally drive past the neighbouring MP’s office on your route. (Or it least this was true prior to 2019 – haven’t checked where the new office is).

        It’s not an intuitive result so it’s not easy to explain to voters. I should bust out my old Urban Operations Research textbook from undergrad (I was trained as a transportation engineer) to quantify this.

        The median district size in Canada is also just 3,000 km2. In Scandinavia their districts range up to 100,000 km2. So there’s more room for larger districts than people realize I think. Greenland (2,000,000 km2) also scrapped its geographic constituencies in the 1990s and hasn’t seen a need to go back. As one party official told me “Earlier, until 1999, we were also obsessed with geographical representation. The problem with that are, that the elected apparently had a tendency to mind that; which part of the country you understand you are representing, and not so much the common good for all citizens.”


        • I think the rationale for having somewhat local districts is the importance of ongoing f2f encounters for building community which is important for activism. Shared ideology alone doesn’t bend the arc of history, in my book, people valuing their neighbors and themselves in part by participating in reworking the rules that govern them is more important.


  5. While this may be an example of irresponsible opposition, it’s seemingly another trend point for Anglosphere democracies in recent years.
    The NZ 2017 election saw very similar lines being used by the (Conservative-equivalent) National Party, which have continued to be used by anti-vaccine conspiracy theorist groups even this year. I don’t think it’s an accident that this comes after Trump supporters claimed the US Presidency was “stolen”.
    It’s known that English-speaking right wing parties often share campaign advisors and tactics – this would seem to be a recent, and disturbing, trend among those parties.


      • But let’s not take it to be too new a phenomenon. Earlier, on Twitter, someone suggested that back in the days of Harper such harsh charges would not have happened. I responded:

        Back in those halcyon days of responsible right-wing politics they would only shut down parliament to prevent the indignity of having the other parties vote them out, while decrying the “coalition of socialists and separatists.”

        Perhaps that is not as bad as the charge of “anti-democratic” we heard last week. But it was still arguably an abuse of the prorogation power and rhetorically pretty scorched-earth-ish. And it worked.


    • This is why I’m hoping to build on the approach taken by the creator of the DMP model. I have a Canadian friend who lives/works in the US but may help me to attract others to campaign for a parallel system approach that would have only 4-seat ridings, plus 1 extra seat per province (including 1 for the 3 territories). It would be about 3/4ths PR, 1/4th majoritarian and have just one vote per voter, similar to the rule used in Italy, with each provincial party specifying a vector 3-4 candidates per riding.
      Then, the outlier voter would have a chance at electing an MP to the 3rd of the 3 PR seats in their riding, or being the swing median voter for either the 4th seat in their riding or the at-large seat in their province.


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.