This is what I submitted to the committee. Who knows whether it will be accepted as a brief? I doubt it will have any influence on the committee's decision, but what is electoral reform other than tilting at windmills.
Brief to the Special Committee on Electoral Reform
Paul Nijjar
Kitchener, Ontario
2016-10-05
The List-Free Mechanic
I endorse Mixed Member Proportional representation (MMPR) as a good electoral system for Canada. It does a good job of maintaining local representation with good proportionality.
Unfortunately, discussions of MMPR often get bogged down in discussions of how to fill top-up seats (aka "list seats" or "party seats"). Who decides the makeup of the lists? From where do the candidates come? People spend a lot of time arguing between two approaches: closed lists (where parties publish lists of candidates before the election) and open lists (where voters have the option of choosing or ordering candidates for the list). Neither of these options is ideal. Voters don't like closed lists because they dislike giving political parties more power over candidate selection. Open lists increase the cognitive burden on voters, who then need to evaluate several party candidates when voting.
In my opinion the closed-list vs open-list debate is a false dichotomy. Historically, these two mechanisms were used to fill list seats in list-PR systems. However, adding local riding elections as we do with MMPR creates a better alternative: a "list-free" (also known as "implicit list", "best-runner-up", "best-loser", or in Hutcheon and Tomek's submission, "repechage") mechanic. My primary plea in this brief is to beg the committee to consider the list-free mechanic for any voting system to which it might apply, including but not limited to MMPR.
The List-Free Mechanic
Briefly stated, the list-free mechanic compares candidates running for the same party within a region. It orders the candidates according to how well they did in the region. First, candidates who won riding races are eliminated from the top-up pool. Then top-up seats are filled according to how well candidates did within their ridings.
This is easier to understand with an example. Consider the five ridings of Kitchener Centre, Kitchener-Conestoga, Waterloo, Cambridge, and Kitchener South-Hespeler. For this example, pretend that these five ridings form a region from which top-up MMPR members of parliament will be drawn, and pretend that proportionality dictates awards the Conservatives a top-up seat.
In my riding of Kitchener Centre, the results were as follows:
- Raj Saini, Liberal, 48.78% of the vote
- Stephen Woodworth, Conservative, 30.36% of the vote
- Susan Cadell, NDP, 16.60% of the vote
- Nicholas Wender, Green, 3.05% of the vote
- Slavko Miladinovic, Libertarian, 0.99% of the vote
- Julian Ichim, Marxist-Leninist, 0.21% of the vote
In this case Stephen Woodworth received 30.36% of the vote in his riding. We now compare Woodworth's share of the vote against other Conservative candidates in the region:
- Harold Albrecht, Kitchener-Conestoga, 43.29% of the vote
- Gary Goodyear, Cambridge, 38.65% of the vote
- Marion Gagne, Kitchener South-Hespeler, 36.68% of the vote
- Peter Braid, Waterloo, 32.28% of the vote
- Stephen Woodworth, Kitchener Centre, 30.36% of the vote
Harold Albrecht won his riding, so he does not qualify for a top-up seat. Therefore the first Conservative top-up seat for this region would go to Gary Goodyear. If a second top-up seat was awarded, it would go to Marion Gagne.
Note that I am not advocating that these five ridings ought to form a region, or claiming that election results would be identical under MMPR as they were under FPTP. But I hope this example illustrates the list-free mechanic clearly.
Advantages
As the committee has seen, many submissions to the ERRE consist of voting nerds (and a few regular Canadians) passionately advocating for their pet voting systems. Why should the committee pay attention to mine?
The system is in production: MMPR with the list-free mechanic has been used in the German state of Baden-Württenberg for decades. Exciting untested ideas from other submissions should not be dismissed out of hand, but using a well-used and well-understood system is a strong point in favour of the list-free mechanic.
Canadians like the mechanic: Although I cannot point to absolute proof, time and time again I have observed people who dislike MMPR change their opinions when they learn about the list-free mechanic. It is not difficult to understand why: many Canadians have seen good candidates fail to get elected because they lost close FPTP races or ran for small parties. List-Free MMPR addresses this.
As further evidence, consider the significant number of briefs that explicitly or implicitly use the list-free mechanic in their proposals. Here is a sampling based on other submissions I have read:
FalveyDenis-BlackPeter-RiserMatt-TrentJohn-e.pdf and KirbyJP-e.pdf use the list free mechanic to implement a form of List PR
HermistonRoss-e.pdf, SmithTrevor-e.pdf, and HutcheonDavidA-TomekJennifer-e.pdf propose forms of MMPR with the list-free mechanic.
McNeilDonna-e.pdf : proposes mixed-member majoritarian with the list-free mechanic.
Several other submissions complain about closed lists, or mention the list-free mechanic in some way (eg HornDanielK-e.pdf, HowattJulian-e.pdf dislike closed lists, and RawlsDon-e.pdf mentions list-free MMPR even as he advocates for AV).
Voters are incentivized to vote honestly: one criticism of MMPR is that the riding vote does not matter much. Like FPTP, most riding votes are wasted. The list-free mechanic puts this valuable information to good use. In turn, voters gain incentives to vote honestly, even for candidates with no chance of winning riding seats.
The ballot stays simple and accessible: The ballot stays as simple as the usual MMPR closed-list two-checkmark system, and the cognitive burden for voters stays low.
It is easy to tabulate results: Tabulating riding results is easy and auditable. Allocating and filling top-up seats can be done as soon as all riding results are in.
Dual-candidacy is mandatory: One problem with closed-list MMPR systems is that small parties stop running local riding candidates. The list-free mechanic helps here, although it does not fix the problem entirely. (My preferred solution is to allow party votes only in those ridings where parties field a local candidate.)
Candidates might campaign differently: It is disheartening to attend an all-candidates meeting and find that some local candidates did not bother showing up. This is a consequence of our emphasis on parties and party leaders. Under the list-free mechanic, it becomes more important for candidates to campaign as individuals.
There are no safe seats: There is no guarantee that unpopular local candidates will win seats. Even party leaders will have to work for their seats. I used to think this was a bug; I now see it as a feature.
Disadvantages
No voting system is perfect, and the list-free mechanic has its flaws. In this section I will sabotage my submission by being honest about them:
Seat exhaustion: It is theoretically possible for parties to run out of candidates before they fill all allocated seats. This is mostly a problem when parties are allowed to collect party votes in ridings where they don't run local candidates.
Who do top-up MPs represent?: Do those MPs who become elected serve only the riding in which they run, or do they serve larger regions? If they serve larger regions, do they become "more important" than their single-constituency counterparts? It is important to make the answers to these questions clear.
Riding votes can be split: This might be the greatest practical drawback of this mechanic. If there are many popular candidates in a riding, they might split the vote such that only the constituency winner gets a seat. This limitation probably affects large parties more than small ones.
There are fewer guarantees that more women and minorities will be elected: Closed-list systems supposedly do a better job of electing women and other underrepresented groups to Parliament than systems where voters influence list order. I believe that electing a wide diversity of candidates is important; I also believe that the disadvantages of closed-list systems far outweigh this advantage. I also suspect (although I do not know) that the list-free mechanic will work well for electing underrepresented groups provided that members of those groups run for office.
Inter-party conflict may increase: Because election candidates within a party compete for top-up seats, there could be shenanigans around allocation of party resources to different candidates. I have no good solution to this problem.
The number of ridings probably changes: Most proportional systems require either increasing the number of MPs or making ridings larger. Unfortunately Canadians dislike both options. Maybe they hold some weird belief that we have exactly the right number of MPs now. I reject this belief, and would have few problems either increasing or slashing the number of MPs, but there is little question that this is a politically contentious point.
Independents are disadvantaged: In principle independents can compete as candidates in ridings, but they do not benefit from the party vote. One solution for this is for independents to form an "Independents Party," but this does not solve all the problems.
Overall I feel that the advantages of the list-free mechanic far outweigh the disadvantages, but at the end of the day the committee's feelings matter more than mine.
Guiding Principles
Let us evaluate the list-free mechanic in light of Minister Monsef's guiding principles. When ambiguous, assume that the mechanic is used with a proportional system such as MMPR.
Effectiveness and Legitimacy: When used with a proportional system, distortions are reduced between popular vote and percentages of seats. Most voters will help somebody of their preferred party get elected. Specifically to the list-free mechanic in MMPR, voters have greater incentives to vote for their preferred candidates honestly.
Engagement and Participation: Smaller parties are likely to exist, win seats, and keep bigger parties honest. These small parties may propose policies that appeal to underrepresented groups (eg Youth parties, Aboriginal parties) better than parties do now.
I do not feel that this mechanic has a large effect on increasing voter turnout, but I feel this is true regardless of the electoral system. Mandatory voting increases voter turnout, not electoral systems.
Accessibility and Inclusiveness: This mechanic does not make it easier or more difficult for those who have difficulties accessing the current voting system. However, using the list-free mechanic keeps the ballot simpler than a ranked-ballot or open-list system, which reduces cognitive load.
Integrity of Voting Process: Tallying votes remains fairly straightforward and can be audited easily.
Accountability of Local Representation: This is where the list-free mechanic shines. All MPs will face the electorate in some way. Unpopular candidates from popular parties become vulnerable, and popular candidates from unpopular (or smaller!) parties can be elected.
Other Acceptable Voting Systems
Overall I support most forms of proportional representation. I prefer list-free MMPR, but would be okay with (in rough order) Fair Vote Canada's Rural-Urban Voting, Single Transferable Vote, open-list MMPR, Sean Graham's Dual-Member Proportional system, or even closed-list MMPR.
I would not be thrilled with a Mixed-member Majoritarian system, but I could tolerate it. Stephane Dion's P3 system would be even more difficult to stomach.
I largely oppose Alternative Vote and First Past the Post. Fractional voting systems are interesting but depend upon parties winning riding seats.
Mandatory Voting
I reluctantly and weakly endorse mandatory voting, under the following conditions:
- Anybody should be able to register as a non-voter before the election. This would accommodate those who oppose voting on conscientious or religious grounds.
- All ballots should include a "none of the above" option explicitly on the ballot, so those who are fed up can express their dismay explicitly. (Declining one's ballot is insufficient because it is obscure.)
- Considerations should be made to the fact that poor people have a more difficult time voting and staying on voter lists than others.
- A fine should be used to punish non-voters (as opposed to an incentive rewarding voters). Humans are loss-averse, so fines work in ways that incentives do not.
The primary reason I support mandatory voting is because political parties systematically ignore certain demographics during election campaigns. Parties do not canvas poor neighbourhoods (because "poor people don't vote") and they ignore safe seats in favour of swing ridings. Mandatory voting might incentivize parties to pay more attention to these demographics. If it doesn't, my support for mandatory voting diminishes significantly.
Online Voting
Many Canadians support the convenience of online voting, but many computer professionals oppose it, because they understand computers and know what security nightmares they are. This alone should give the committee food for thought. The so-called "robocall scandal" of 2011 illustrates that there are forces willing to subvert elections, and allowing online voting increases the attack surface significantly.
Estonia has embraced online voting thoroughly, but even their system has problems (for example, see: https://estoniaevoting.org/). That said, Estonia has addressed many common concerns around e-voting, and if the committee wants to pursue this option they should learn from what the country has done.
I remain concerned that by implementing online voting we jeopardize several things:
- We lose the secret ballot, because family members/workplace leaders can coerce those under their control into voting in particular ways. There are ways to mitigate against this (again, see Estonia) but unlike voting booths we cannot ensure that online voting preserves the secret ballot.
- As electronic voting becomes more popular I worry that those who cannot/will not vote electronically (eg those with flaky internet connections) will be left out.
It is not clear that online voting helps with voter turnout. As reported by the CBC, the City of Guelph enjoyed an 11% increase in voter turnout, but the City of Cambridge only had a 1% increase. Meanwhile, Kitchener (which did not have online voting) had a 2% increase.
Elections Manitoba has been conducting surveys of voters and non-voters in their province. They find that although 59% of non-voters said they would be more likely to vote if online voting was available (p. 29), in practice few non-voters cited inconvenience for the reasons they did not vote. 2% claimed that the voting locations were inconvenient, and 1% said polling stations closed too early. (p. 12). In contrast, 20% of non-voters cited reasons related to disillusionment, and 10% said they did not know enough about the parties or issues (p. 12). To me, this suggests that online voting does not address the problems we want it to solve. Being able to vote more conveniently does not make learning about the election any easier.
Having said all this, I acknowledge that online voting could make voting easier for those with mobility challenges. But I feel it is a lot of expense and risk for little payoff, and would like to see accessibility concerns addressed in some other way.
Other Comments
Unlike many other proportional representation advocates, I do not endorse full proportionality at any expense. Regional and urban-rural distinctions are important in Canada, and our voting system should take these into consideration even if it makes the system less proportional.
I believe the per-vote subsidy should be reinstated. This small change made a large difference in voter incentives under FPTP.
There are many details to consider when designing an MMPR system. My preferences are:
- For the list-free mechanic, rank candidates within a party by percentage of the vote, not number of votes.
- Allow overhang seats.
- Set thresholds fairly low (2%?) since regional boundaries will reduce proportionality.
- Use a two-vote system (voters choose a party and a local candidate) rather than a one-vote system (where voters choose a local candidate, and the party is implied).
Although not a question directly considered by the committee, I endorse lowering the voting age to 16.
It is disingenuous to pretend that Canadians primarily vote on the basis of local candidates, as opposed to parties or party leaders. The argument that FPTP is fair because we vote for individual candidates and not parties is similarly disingenuous. However, I agree with the underlying criticism that political parties are too powerful.
I agree with Denis Pilon's arguments concerning referenda. To those not satisfied by Pilon's arguments, I offer the following thought: the next election will serve as a referendum for these electoral changes. If Canadians vote into power parties that pledge to return the country to FPTP, then it only requires an act of Parliament to reinstate FPTP.
I am disillusioned with fixed election dates. They do not dissuade governments from calling early elections, and they result in a lot of pre-campaign campaigning.
It will be disappointing if the committee's final report breaks down across party lines. If it does, expect voter cynicism in electoral reform to increase yet further.
Summary
- The committee should adopt the "list-free" mechanic for voting systems that select candidates from lists.
- Nobody likes mandatory voting but it might be warranted.
- Online voting is probably inevitable, but it is a bad idea.