One Vote MMP
Whenever some government (like that backstabber Trudeau ) dangles the promise of electoral reform in front of the Fair Vote zealots, the Fair Vote zealots immediately rips themselves to shreds with internecine warfare about "the best voting system". Everybody has their preferred system, and everybody advocates for their preference to the exclusion of everything else. Backstabber Trudeau only supported Alternative Vote, which is why he wasted all of our time and energy in 2015 (going so far as to appropriate FVC slogans) only to scuttle the enterprise when it was apparent the Fair Vote zealots would only accept some kind of proportional representation, which he (and nearly all Liberals) hate with a vengeance.
In that light, the worst thing a Fair Vote zealot could do for the cause is to advocate yet another voting system, which is exactly what I am going to do here. I would like to disclaim that it is not my favorite voting system, but it is an interesting one with some solid advantages over other proposals.
This voting system came to mind when reading the Conservative Party platform with respect to electoral reform. It reads:
- Electoral Reform
The Conservative Party believes the discussion of possible changes to the electoral system is valuable in a healthy democracy.
In reviewing options for electoral reform, we believe the government should not endorse any new electoral system that will weaken the link between Members of Parliament and their constituents, that will create unmanageably large ridings, or that will strengthen the control of the party machinery over individual Members of Parliament.
A national referendum must be held prior to implementing any future electoral reform proposal.
The Conservatives think they are being very clever here. They carefully word this resolution to exclude every form of proportional representation:
"weaken the link between Members of Parliament and their constituents" is intended to exclude any form of MMP or pure PR that uses party lists.
"that will create unmanageably large ridings" is intended to exclude STV, which achieves proportionality by combining several ridings to form multi-member constituencies. It is also a dig against forms of proportional representation that reduce the number of constituent seats to add some proportional seats, because we "don't want too many politicians".
"strengthen the control of the party machinery over individual Members of Parliament" is another dig at party lists.
All of this is nonsense, of course. The Conservative Party (and in particular the leader) has an iron grip on CPC members of parliament. Every candidate has their nomination paper approved by the centralized Party membership. The link between MPs and their constituents is nearly nonexistent -- Conservative candidates don't show up for debates, and their votes are dictated and whipped by cental Party leadership, to the exclusion of whatever their constituents want. There are already unmanageably large ridings but the Conservatives are in no hurry to change those. It's all stinking hypocrisy. But it is also interesting, because those sneaky Conservatives haven't covered all the bases. They have not ruled out list-free MMP.
The gist of list-free MMP is a form of Mixed-Member Proportional voting where there are some fraction of local riding seats (just as today) and then some "top-up" seats to maintain proportionality. So the PPC might not win a single constituent riding, but they might qualify for one or more top-up seats.
Where do these top-up seats come from? In most versions of MMP they are filled according to a "closed party list," which is made up of names that the central party leadership chooses and then fills in order. So-called "open list" systems give voters some influence over the ordering of this list, creating giant ballots that are headaches for everybody. People in Canada HATE closed party lists, so they are a common target for FPTP advocates (including backstabber Trudeau, who uses this argument to justify why Proportional Representation is unacceptable).
However, there is a third option. Instead of filling the top-up seats with an explicit list, we can just pick top-up winners from people who ran for constituent seats. Say the PPC win no riding seats but deserve three top-up seats. Rank each PPC candidate by the proportion of the vote they received in their local races, and then use that to fill the top-up seats. If some of those PPC candidates had won a riding seat, we skip them and go down the list until we reach the top three that did not finish first in their ridings.
As a concrete example, consider the following PPC results (which are not exhaustive, and picked arbitrarily from the CBC vote tracker ). The format is: candidate name, riding name, proportion of vote in that riding, and total votes.
- Isabel Pereira, Parry Sound-Muskoka, 1.6% share, 1048 votes
- Todd Hoffman, Pontiac-Kitigan Zibi, 1.1% share, 673 votes
- Patricia Conlin, York-Dunham, 1.2% share, 901 votes
- Kevin Dupuis, Kitchener-Conestoga, 1.3% share, 786 votes
- Henry Geissler, Haldimand-Norfolk, 0.9% share, 657 votes
- Michael Manchen, Lakeland, 1.7% share, 982 votes
If we reorder by vote share, we get:
- Michael Manchen, Lakeland, 1.7% share, 982 votes
- Isabel Pereira, Parry Sound-Muskoka, 1.6% share, 1048 votes
- Kevin Dupuis, Kitchener-Conestoga, 1.3% share, 786 votes
- Patricia Conlin, York-Dunham, 1.2% share, 901 votes
- Todd Hoffman, Pontiac-Kitigan Zibi, 1.1% share, 673 votes
- Henry Geissler, Haldimand-Norfolk, 0.9% share, 657 votes
So out of this list the first three list seats would go to Michael Manchen, Isabel Pereira, and Kevin Dupuis, in that order. Note that Dupuis wins over Patricia Conlin even though she got more total votes, because Dupuis has a higher share of the votes in his riding.
In an actual election we would have ranked all the PPC candidates, not just six. We may have also done some splitting by region, depending on how we architected the system. But this is the basic idea behind list-free MMP.
This voting system does not weaken the link between MPs and their constituents, because the candidates that win top-up seats are by some definition the most popular in their party, as determined by the electorate.
This does not create large, unmanageable ridings, because we can keep the number of ridings the same and just add 30% more seats as top-up seats.
This does not strengthen the control of party machinery over individual members of Parliament, because as long as the candidates can run in individual ridings (which is as much control as the party machinery has now), the party machinery cannot choose how to fill top-up seats. Admittedly, the party machinery has control of how many top-up seats they win (because that is determined by the proportion of voters nationwide who support them) but that is not additional control over individual members of Parliament.
That is all old news.
The new insight is that a list-free MMP system need not have a standard MMP ballot. A standard MMP ballot gives each elector two votes -- one for a party and one for a riding candidate. A voter might choose to award their votes to the candidate and their party, or might choose to split them (for example, if you support the Green Party but really like the local PPC candidate).
So what is the insight? The insight is that MMP does not actually require two votes. You could have an identical ballot to FPTP, and simply use that to determine party support. Once you know party support across the country, you can allocate top-up seats accordingly. Formally, this system is known as "mixed single-vote".
In some sense, this system is what Fair Vote zealots use already. Whenever we want to show the vast difference between the popular vote and the number of seats won, we tally up the votes for local ridings and pretend that they equal the share of the party vote a particular party would have won. I made many such graphs when I was a Fair Vote zealot, and so has everybody else.
The Proposed System
Here is a summary of the proposed system:
- Use list-free MMP with local ridings as they are and some top-up seats.
- Fill the top-up seats using the "best loser" algorithm above.
- The ballot remains identical to the ballot we have now, which has voters choose one local candidate (most of whom are associated with a party).
In the analysis below, I will compare the advantages and disadvantages of this proposal against two-vote list-free MMP and against FPTP.
Advantages
As compared to FPTP, the system has the following advantages:
- It is a less disproportional system.
- Voters really like the idea that popular candidates who do not win their ridings get rewarded, and that unpopular riding candidates don't get promoted by their parties to top-up seats.
- Parties that do not win individual ridings but have a fair amount of support across the nation get some representation.
- This vastly reduces the pressure for people to vote strategically.
- List-free MMP in general has the huge advantage that people are best off voting honestly, because they will get some representation somewhere. That reduces the coordination headaches of strategic voting (which really does not work, and which makes everybody upset).
As compared to two-vote list-free MMP, the system has the following advantages:
- Since the ballot is the same, there is no possible complaint that the system is too complicated for voters to understand.
- It completely eliminates "decoy list" plays by big parties (although you can easily get rid of these in two-vote MMP by only allowing party votes if there is a corresponding riding candidate on the ballot).
Disadvantages
As compared to FPTP, the system has the following disavantages:
- It means the Conservatives and Liberals do not get all the power, so will never ever be implemented. The major parties hate anything even approaching proportionality.
- This takes away control of MP election from central party machinery, which the parties (including the Conservatives) hate.
- There will be stupid arguments about "too many MPs" if you keep the number of ridings the same and add 30% top-up seats. If you keep the total number of ridings the same then you will fail the Conservative "unmangeably large riding" test. Both of these arguments are stupid (especially given the centralization of power in the Prime Minister's Office) but they will come up all the time.
- Both two-vote MMP and this system create a few instances where a person who finishes third or fourth in a riding nonetheless wins a seat, because they are relatively popular in their party.
- Similarly, in rare cases three or four candidates from the same riding might all be elected to parliament.
As compared to two-vote list-free MMP, the system has the following disadvantages:
- In two-vote MMP there is a clear distinction between supporting a party and supporting an individual candidate. In FPTP this is all confused. If you vote for your local PPC candidate, does that mean you support Maxime Bernier? That you support the PPC? Or that you support the local riding candidate? Two vote MMP splits up one of these distinctions. One vote MMP makes things ambiguous again.
- Historically the one-vote system was initially tried in Germany. Then it was abandoned. I am not sure exactly why, but this is evidence that something did not work.
Outstanding Questions
- In Baden-Wurttemberg, how often do top-up riding members end up winning riding seats in subsequent elections?
- Why does Baden-Wurttemberg run two nominees per party in each riding? Is this necessary?
- What happened in Lesotho?
- How do you prevent MMP systems from devolving into parallel voting systems, as happened in Thailand and some other jurisdictions?
Overall
I think this system is now my second-most preferred system for electoral reform. The list is now:
- Two-vote list-free MMP
- One-vote list-free MMP
- Single Transferable Vote (STV)
- List-free parallel voting
and that's about it. I do not like pure PR systems. I do not like either open or closed-list MMP. I do not like Alternative Vote, but I am guessing that is what we end up with. Weird hybrids like rural-urban proportional systems are cute, but I do not think they solve the real problem.
In fact as an apostate Fair Vote zealot, I mostly think that proportional representation does not solve most of the problems we actually want. I also no longer think that pure proportionality is actually the goal I want as a voter. My priorities are:
- If I cast a ballot do I usually get some political representation for it? Or am I as doomed as a Liberal voter in Alberta?
- If a political party wins the most votes overall does it get to form government? Or do we get wrong-way winners, the way Backstabber Trudeau benefited from wrong-way winners in 2019 and 2021 . Even though I overall preferred Backstabber Trudeau to the corresponding CPC governments, the wrong-way winners were still an injustice.
- Do underrepresented groups (including rural votes, Quebec voters, indigenous voters, and (sigh) Alberta voters) get a reasonable share of power based on their concerns? This is a very important consideration in Canadian politics, and it does not map to strict proportionality well at all.
Proportional voting is one way to solve most of these problems, but it clearly leaves some groups (Quebec voters, rural voters) in the lurch. I am happy to accept some disproportionality so that these groups have a more fair share of political power as compared to champagne-sipping coastal elites like me.