Downs? Duverger? Both? Canada 2015

Returning to the theme of my pre-election entry, did the Liberal majority of seats result from the “median voter’s” swing (Downsian competition) or from strategic desertion of the third party (Duvergerian tendencies)? Of course, it can be both.

There was clearly a late swing from the NDP to the Liberals–maybe a little more than a couple of percentage points. This looks like strategic voting, not wanting to waste the vote, ensuring defeat of the incumbent by voting for the less-preferred challenger. In other words, Duverger. In fact, this process had been going on for a while, with the NDP slide and Liberal rise being quite correlated in the CBC poll tracker. Yet for most of the final two weeks or so, it seems as if the Liberals might have gaining mostly at the expense of the Conservatives. That still looks Downsian to me: the swinging of votes from incumbent to the more “centrist” challenger. I will leave it to others to define where the median voter lies in Canada, but it seems that he or she prefers a red hue (red being the color of the Liberal party).

One way in which the outcome is not Duvergerian is in the Liberals’ success in Quebec. The party won many more seats there than expected. In fact, it won 40 of the 78 seats in Quebec (51.3%). I have not located a provincial level voting breakdown (anyone?), but on CBC last night it was reported at around 35%. It was further reported that several ridings (districts) were won in the province with just over 30% of the votes. That is not especially “Duvergerian”.

Of course, the presence of a third nationwide party with around 20% of the votes and 13% of the seats is also somewhat of a challenge to Duverger’s Law. This is one of the better showings for a third party in Canada in recent decades.

It is also noteworthy that the mean district-level effective number of parties, by votes, was around 2.7 in this election, hardly changed from 2011, although with more dispersion around that mean (thanks to Ko Maeda, personal communication, for the stats).

A few basic facts about the national level result…

The advantage ratio (percent seats over percent votes) for the Liberals was 1.38. That is very close to the mean (1.35) and median (1.33) of a dataset of 210 FPTP elections I assembled and have done comparative work with.

The size of the the third party, as a percentage of all seats, is at about the 90th percentile. The third party’s vote share is somewhere between the 75th and 90th percentile.

The effective number of parties by votes is 3.33; by seat-winning parties it is 2.50. The figure for votes is well above the mean for FPTP parliamentary systems, which is 2.76; the mean for seats is 2.15. In both of these indicators, this election is again between the 75th and 90th percentiles.

It seems from these figures that Canada still offers a fairly objective case for the adoption of proportional representation: its party system is more multiparty than what is typical for FPTP electoral systems. Whether this election makes such a change more, or less, likely is unclear. On the one hand, majority parties aren’t often known for their enthusiasm about PR. On the other hand, the Liberals were just elected with a manifesto commitment to electoral reform, although not specifically PR.

19 thoughts on “Downs? Duverger? Both? Canada 2015

  1. Antony Green’s preliminary thoughts on how AV would have altered the result. The punchline is:

    The contests where preferences might have changed the result were in Quebec, or a smaller number of critical three-way contests between the Liberals, NDP and Conservatives, mainly in urban seats in western Canada.

    The NDP won 25 seats with support under 35%, it having been the party most squeezed by multi-party contests, especially in Quebec.

    At some point in the future I’ll take a closer look at the contests and possible impact of preferences. However, on my first analysis, the impact of preferences would have been in line with Australian experience, perhaps changing the result in one and ten ridings, and those changes being concentrated in ridings where the winning candidate recorded a low percentage vote share.

  2. What would have happened to the election result if Canada had used a French 2-round system a la the French National Assembly? Would the result be different especially if their are two rounds? The problem with that system is the cost of having 2 elections with sometime inbetween.

    • You also get a significantly different cohort of electors at the 2 rounds. In 2012, 35,883,209 people voted in the first round but only 34,861,353. voted in the second round. Excluding 2.85% of first round voters is a pretty unhappy situation. The old saw that the French voted their hearts in the first round and their heads in the second only indicates that the second round cohort is more conservative than the first round cohort. This second ballot was originally adopted as a way of ensuring precisely that result.

      Moreover, because the electors cannot converge on a candidate, as is possible under AV, you get bizarre situations like the 2002 election where Chirac achieved 19.88% in the first round and blew out to 82.21% in the second round because electors who voted for candidates other than Chirac or Le Pen in the first round were conscripted by the system to vote for Chirac in the second round because Le Pen was so widely unacceptable.

      The various AV ‘alternatives’ like supplementary vote have the same problems as the second round.

    • I agree with Alan here. I can’t really see any substantial advantage for a two-round system over AV, and quite a few disadvantages (higher cost, potential lousy top two). However, when you say the second round cohort is more conservative, does that mean supportive of the political right, or supportive of the status quo? The 2007 French legislative election was interesting, because it showed that voters can change between the first and second rounds to reduce a potential large one-party majority. The second round seems to prevent radicalism (it’s anti-National Front and anti-Communist) rather than supporting either left or right.

      • The second ballot favours the side is that is more unified. In France that has historically been the right. There was no prospect of electing a president from the left while they remained sharply divided between the Communists, Socialists and a number of minor parties. Really it is a temporal gerrymander.

      • How is that, Alan? Surely plurality is the system that favours the side that is more unified. Unless you’re purely comparing AV and majority-plurality TRS, I don’t really understand your claim.

  3. This is one of the better showings for a third party in Canada in recent decades.

    Is it really? Let’s look at the last thirty years of elections, and the proportion of seats the third party received:

    * 2015: NDP, 44 / 338 = 13.0%
    * 2011: PLC, 34 / 308 = 11.0%
    * 2008: BQ, 49 / 308 = 15.9%
    * 2006: BQ, 51 / 308 = 16.6%
    * 2004: BQ, 54 / 308 = 17.5%
    * 2000: BQ, 38 / 301 = 12.6%
    * 1997: BQ, 44 / 301 = 14.6%
    * 1993: Reform, 52 / 295 = 17.6%
    * 1988: NDP, 43 / 295 = 14.6%
    * 1984: NDP, 30 / 282 = 10.6%

    The last election’s 13% doesn’t strike me as particularly high, compared to other results. What’s actually surprising is that this is the first time in 27 years that the NDP is in third place!

    • I think I mean to say that I meant the third party as defined by votes, given the context was whether Duvergerian pressures had kicked in. I also almost certainly meant to say “third national party”, leaving the BQ with its FPTP-fit regional concentration out of the mix. Sorry for not being clear.

      • Ah, yes it does make sense to look at it by vote percentage. Sorry about my assumption earlier. Here’s the list:

        * 2015: NDP, 19.7%
        * 2011: LPC, 18.9%
        * 2008: NDP, 18.2%
        * 2006: NDP, 17.5%
        * 2004: NDP, 15.7%
        * 2000: PC, 12.2%
        * 1997: PC, 18.8%
        * 1993: PC, 16.0%
        * 1988: NDP, 20.4%
        * 1984: NDP, 18.8%

        The latest NDP result is on the high side, but not by much—there is a clear but slow upward trend from 2000 until now. It’s surprising to me just how many votes the Progressive Conservatives were still getting in the 90s, though!

  4. Will there be a referendum on this issue or will it be enacted by Canada’s parliament? Wow, that would be amazing if the Liberals enacted PR when they have an absolute majority.

      • Unless I am mistaken, the manifesto does not say there will not be a referendum, but it does not say there will be one.

        There is a long way from where we are today to the possibility of a formal proposal for a new electoral system, but if there is such a proposal, I would find it highly unlikely that it would be implemented without a referendum. I also find it, at the moment, unlikely that such a referendum would pass, but this comes with very big caveats that we can’t know right now what a referendum campaign and public education effort might look like.

        Despite all that, the manifesto itself is pretty clear: this would be the last FPTP election. So if they renege, it will mean the calculation is that there won’t be enough voters who care enough about the issue to be costly at the polls in 2019. Which is pretty far off…

      • At a post-campaign panel on CPAC, one of the Liberal campaign managers (?) fought pretty hard not to commit to a referendum. I suppose if they had the support of some of the other parties, they might be able to pass a bill directly without it looking too self-serving.

        On the other hand, the NDP campaign manager made sure to mention “proportional” every time she talked about the issue, so I assume the NDP wouldn’t support just AV, maybe not even STV.

        Hopefully if a referendum does happen, at least some of the political parties actually come out in support of the proposal. It’s pretty frustrating when the parties initiating the referendum are so half-hearted about it.

  5. Pingback: Trudeau promise #1: Electoral reform | Voting in Canada

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 )

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.