Different electoral system, different coalition

If you want to discuss this intelligently, stop using the term “proportional representation” or “PR” when you mean “one of a variety of more party-proportional systems”. Both STV and MMP are more party-proportional, but they are radically different in the way they survey the public’s views of their leadership. There is also instant-runoff-voting (IRV) in which one simply ranks candidates in existing single-member districts, which is also useful for Mayor or referenda where one might be, say, choosing between voting systems. The simplest option federally is to do exactly what Harper has proposed for the Senate and what his own Conservative Party of Canada did to elect him: just adopt an instant-runoff ballot for the House in single member districts identical to the ones we now use. This will end the “vote splitting” but it will not elect as many minor party members as STV and certainly nowhere near as many as MMP. Given that MMP was soundly rejected (for good reason) by PEI and Ontario, and STV was embraced by 77 of 79 districts and got nearly 60% support in BC (and would have passed under a reasonable formula) we should also consider IRV to be a first step to a possible reform towards STV federally. By no means, in either Senate or House, would it be acceptable to have votes in BC determine outcomes in NL. But that’s not the reason voters rejected MMP, it seems they don’t like parties choosing cronies to put on “party lists”.

If you want a comprehensive analysis that leads to the same conclusion, that no more than a few party-determined seats (10-15% would be a lot) would ever be tolerated in Canada, here’s one, by myself based on an earlier version by Julian West.

== Voting system values ==

(Voting system values and related goals and guarantees of general electoral reform in Canada)

The ideal or fundamental way to design a voting system is to agree first on what values and principles are appropriate to incorporate. Within the Canadian voting reform community, there is remarkable unanimity on what values are important for Canadian voters. The constitution of Fair Vote Canada expresses these values as follows:

* 1) broad proportionality,
* 2) extended voter choice,
* 3) stable and responsive government, and
* 4) maintaining a link between representatives and geographic constituencies.

This list resembles the list of requirements issued by the British government to the Jenkins commission. It was first proposed in the Canadian context by Nick Loenen, and then adopted, with minor modifications, by Fair Voting B.C. and Fair Vote Canada. A similar, but slightly longer, list was used by the New Zealand Royal Commission and was in 2005 adapted by the Law Commission of Canada.

Any rigorous literature search will reveal broad agreement about these values. Politicians and electoral reformers on the left and the right, and from all political parties large and small have accepted them with some enthusiasm; They have never been seriously attacked. At Fair Vote Canada’s founding meeting, they were adopted by a roomful of grassroots members, with consensus being reached after one member proposed the inclusion of the words “and responsive”, having articulately questioned whether “stability” alone is the actual goal. However this list is not itself complete and doesn’t take into account experiences of Canadian electoral reform advocates in actual referenda.

While lawyers and professional politicians may agree only on these four broad values, their view is not the only view we must consider. The public has political instincts too.
Additional values are held, and guarantees expected, by the public. These are extant in academic study of the role of governments (1a, 2a, 3a, 4a) and in actual experience with proposed electoral reforms in Canada (5, 5a, 6, 7, 7a ) and the US (6a) and especially counter-intuitive results (6, 7a). The degree to which any change requires learning and creates errors also requires us to minimize the disruption of any changes (per 6b).

Let’s consider boundaries. Among political scientists, there is also broad agreement that government has the primary responsibility to manage the relationship between the persistent lands and waters and ecosystems on which we live, and the human infrastructures built on it. When dealing with these very fragile relationships, bioregional borders are recognized as an optimum that is both stable and responsive as a population grows and natural systems come under stress: water, soil, invasive plant species, effectively move only in one direction: down via the watershed to the rivers, lakes and sea. It is probably only these borders that can provide long term stability, responsiveness and objectivity. Many experts in international relations have commented on the effects of arbitrary borders drawn to separate people of similar lifeways and dependent on similar ecosystems, and the potential to create chaos if policies on how to manage natural ecosystems vary too broadly. For these and several other reasons, Canadian municipal borders remain relatively stable compared to federal and provincial districts. A system capable of dealing with fixed borders and changing populations also responds instantly to any such shifts.

The expert and public opinion suggests a few specific guarantees, each of which extends one of the basic four principles and all of which are satisfied by at least one voting system [1]:

** 1a) instant adjustment to ensure proportionality so every vote counts exactly the same without districts having to change
** 2a) a fair chance for community-based independents – including municipal politicians – without strong (or any) party ties
** 3a) aligning municipal divisions more closely to the ecosystems they manage and to stable federal/provincial districts
** 4a) guaranteed fixed and stable divisions of political responsibility for ecosystems protected and used in common

Three further implicit values invariably become clear considering specific systems, each with a related guarantee demanded by at least some of the advocates of that value. These concern election timing, process, public input into the actual persons who vote in legislatures.

* 5) a relatively predictable election schedule so that small parties or those choosing leaders are not unduly disadvantaged
** 5a) fixed election dates guaranteed by statute that can only be over-ridden by a specific protocol that delays the writ drop

This constraint is likewise more easily satisfied by a system in which seat shifts are generally proportional to the popular vote rather than being wildly volatile with small shifts in that vote in specific districts. The motivation to disrupt and dissolve legislatures is almost nil since no party is likely to gain a large number of seats by doing so at a particular time, and cannot benefit from vote-splitting among its rivals.

* 6) simplicity insofar as an ordinary member of the public can learn and practice the counting system with no expertise
** 6a) zero tolerance for any system that does not guarantee a paper trail, paper recounts, or relies on electronic storage
** 6b)backwards compatibility of the ballot so that those familiar withi FPTP voting systems can participate without spoilage

The nearly universal resistance to “e-voting” among neutral non-industry-allied experts is evidence that guarantee 6a is very important to reformers. Many US states have passed laws requiring ballot paper trails and forbidding reliance on any system that only experts can verify. While arguments regarding the complexity or simplicity of a voting system are often raised with the assumption that simpler is better, it is absolutely clear that Canadians do not understand how FPTP system works nor how it leads to such vastly disproportional results. The simplicity of the ballot itself, in other words, is not necessarily as important as the simplicity of candidate selection, counting systems and understanding how to change one’s representative. The Canadian experience has also established that simpler systems, well understood, such as MMP, are roundly rejected in favour of more complex systems, such as STV, that satisfy other constraints, notably this last one:

* 7) the public themselves, not the political parties, makes the final choice regarding who is to be trusted with real power
** 7a) every individual who acquires a vote in the legislature must have been selected by a process open to public input

Failure to recognize this seventh value led Prince Edward Island and Ontario both to propose MMP systems that allocated about one third of all seats to “party list” candidates chosen by the parties and leaders. Both proposals were overwhelmingly rejected by the public almost certainly because of their failure (or perceived failure in the Ontario case) to satisfy this extremely important public trust constraint. American political parties invariably guarantee 7a at least for registered members of their political party. In Canada only party leadership races are subject to any public scrutiny [2].

Any voting system should be assessed against all seven of the broad and all seven specific values. Informed advocates will certainly intensely scrutinize any system that does not sastify or potentially satisfy all seven values and eight specific goals and would certainly oppose, for instance, systems like closed-list MMP implemented electronically (violating 1a, 2a, 3a, 4a, 6ab, 7, 7a). Presenting a simple choice between status quo FPTP and such a closed-list MMP system must be viewed as a sabotage of the reform process, given the sub-40% outcomes in PEI and ON. The remaining options in Canada are single winner or multi-member ranked ballots (IRV and STV) and a mixed STV system with only a few party-proportional seats (such as Craig Hubley’s BSTV+C+L proposal).

Any voting system represents compromises. A report by the Scottish parliament regarding the use of STV in Ireland noted that “the heavy emphasis on constituency casework, faction-fighting between candidates from the same party, a focus on constituency, localist matters in election campaigns and parliamentary work, ‘friends and neighbours’ voting, are all seen as resulting – at least in large part – from the candidate-centred, preference voting of STV” [http://www.scottish.parliament.uk/business/research/briefings-03/sb03-85.pdf]

Whether a constituency or local focus and visible conflicts within parties are a good or bad thing is entirely a question of one’s values. It would appear, however, that the public in Canada errs more in favour of these things than against them, judging by MMP’s failure in ON and PEI.

[1] A chart of how each system satisfies the guarantees is available from craighubleyca@yahoo.com

[2] Accordingly Hubley’s BSTV and B5AV permit only leaders and simultaneously elected deputies to reach the legislature by means of the “+L” seats. And that would only be permitted if the most popular elected member of that party chooses to step aside, and if the party’s advertised constitution requires them to and their political advertising made clear that this might occur. It could be required also to hold a ratification by-election to clear this substitution with the public, though that would possibly be onerous.

CANADA NEEDS AN ELECTORAL SYSTEM THAT WILL BUILD COMMUNITY IN CANADA

Canada is presently experiencing a serious Parliamentary Crisis that is caused by our archaic system of electing our members of Parliament.
Canada’s next election should be a meaningful election in which voters can vote for whom and what they believe in, with confidence that their vote will affect the formation of Parliament. In the 2006 and 2008 elections, less than 50% of electors could point to an MP that their vote helped to elect. Using the newly proposed electoral system, over 90% of electors would.
If Canada had an effective electoral system, minority governments would be almost a certainty. Prime ministers and their governments would need to learn how to work collaboratively for the good of a more inclusive Canada. As in this crisis, if the focus was power, and the Government could not work collaboratively with the confidence of the House; a coalition government could be formed to carry out the will of the voters in the previous election.
With a minority government working collaboratively, bills like those that caused this crisis would not be brought forward. If Canada now had a majority government, those bills that would have reduced the feeling of community in Canada would have passed. Under Canada’s present system, the majority government would almost always be a false majority, with less than half the votes but more than half the seats (usually about 40% of the votes).

The system proposed on the following website is relatively simple and would be easy to implement in Canada. Ridings would be twice as large. Keeping the same number of seats in Parliament, half the seats would be riding seats and half would be proportional seats. To keep local representation, the seats would be divided into areas with between four and eight seats. Each area would have its own popular vote for designating the area’s proportional seats. All seats would be occupied by candidates who won their seat in the election.
Because we would have a far more representative democracy, all areas of Canada would be represented by both Government and Opposition MPs. This proposed electoral system would help build community in Canada.
Click on this link for more Info: http://www.electoralchange.ca. At the bottom of the first page, you will find links to both How This Change Could Happen plus a PDF System Presentation.
David Brekke, B.Ed., M.Ed.
Former Federal Returning Officer and
Member of Elections Canada Returning Officers Advisory Committee
Very Concerned Canadian Citizen with an Apolitical Position

Of all the disastrous effects of first past the post none is worse than the way it forces political talent to either join a dominant party or stay out of politics altogether. Who knows how many great Liberals from Alberta, Conservatives from Toronto or PEI, NDPers from Newfoundland, have never been able to get to the legislature because the dominant parties in their region simply offer someone with their beliefs absolutely no path to win?

This country can survive perhaps two more elections with this system before it is utterly eradicated by the way it enables narrow regional interests and systematically disempowers parties that seek strong national presence. The time to act is now, but reviving failed 30-50% MMP schemes won’t work.

Consider:

Two mixed-member-proportional (MMP) systems even less radical than the one David Brekke (electoralchange.ca) is proposing were soundly rejected by ON and PEI voters in 2007 and 2005 respectively, with about 38% of the vote. Far less than the 60% approval threshold usually set for such referenda.

Each of these schemes would have given only about one-third of the seats in the legislature to the parties to allocate according to their lists, as opposed to Mr. Brekke’s one-half. There is no reason to believe that such a system as Mr. Brekke proposes has any chance to succeed at all, and every reason to believe that continuing to discuss schemes of this sort will simply prevent any change. Accordingly it’s not constructive nor practical to be re-introducing failed heavily-MMP schemes to the public – by contrast an incremental instant-runoff-voting reform could be implemented without controversy and (as I’ve explained) Mr. Harper could not oppose it without hypocrisy. The single-transferable-vote scheme, which uses a ballot that is or is nearly identical to the IRV ballot, was supported by nearly 58% of BC in 2005 and by 77/79 of its existing districts. It shares with Brekke’s scheme the division into multi-member districts that preserves regional interests and allows the districts to more closely match municipal and bioregional boundaries, and to change less – meaning less “gerrymandering”. So IRV now, STV later is a viable scheme that ends vote splitting *now*.

If the BC STV scheme passes the 60% threshold in May then this will become the default reform scheme for all other provinces and federally. STV is almost as party-proportional as MMP, and far more representative of public wishes in that it does not permit votes from one side of the province to affect who from the other is elected – also the public must approve of every individual (not “list”) by name. Some very small parties will be under-represented but an analysis of the 2001 BC results showed that had BC-STV been in force the Greens would have had eight seats, including all key critics and leaders, as opposed to the eleven they’d have “deserved” under a fully party-proportional scheme. Parties losing a little influence in favour of the public or regional balance is not a bad thing necessarily: There are other kinds of “proportional” than that to a candidate’s party.

Should STV fail, the door would be open to a more complex scheme that had a very few (10-15%, not 30%, not 50%) party-proportional seats that could be used in a makeup fashion to ensure that all but the smallest parties got in with the same levels of support per member as larger parties. The best of these schemes do not arbitrarily assign a threshold to be represented (and it’s not clear they could since the Supreme Court has ruled that per-vote funding cannot be made conditional on achieving even 2% level of support) but do guarantee that every party above a certain threshold gets elected – that threshold being only statistically predictable in advance but easily observed after the fact. For examples of election calculations of this sort you may contact me directly at craighubleyca@yahoo.com or on Facebook.

However only if BC-STV fails should such compromises even be considered and that only to get MMP fanatics like Mr. Brekke onside. It’s quite likely that introducing *any* party-proportional seats would decrease support for reform. And while complexity of an electoral scheme seems not to matter to the public at all or even be a plus (in BC they arguably voted 58% for a scheme they did not understand, while in ON and PE they voted 38% for a scheme they did understand much better – and objected to for good reasons) there may be a limit to how complex a scheme can be and still be accepted. If the average person cannot be trained in how to count the ballots within say an hour, then, you might start to ask whether the scheme is viable or democratic. (For these reasons, no e-voting scheme without a paper trail can ever be democratic as one is forced to trust experts to tell you whether it is working or not). So 1. IRV 2. STV 3. STV + minimal MMP will remain the most likely order of successful reform, with 30%+ MMP unlikely.

For MMP fanatics (only) I offer the following justification for them to switch to advocating a mixed mostly-STV-no-more-than-15%-MMP scheme that would allow for accountability improvements like fixed electoral borders and the ability of districts to decide to have fewer STV more MMP members, within limits:

An analysis of the most recent Nova Scotia election proved that even if 10-15% of the seats were party-proportional, the only parties that would get them under such a mixed 85STV+15MMP scheme would be those one would expect: four of the six such seats would go to the under-represented Liberal Party of Nova Scotia, and two to the presently-unrepresented Green Party of Nova Scotia. With more than 15% of seats party-proportional, one starts to see allocated seats going to independents and very small “protest” parties – I think it’s obvious that such individuals should usually be directly elected where people know them, rather than rely on questionable affiliations of convenience. Also with no more than 15% of seats allocated in this way it will be persons “almost elected” in particular districts (the Liberals) and persons “privately elected” as leaders and deputies of political parties – both of which are a form of public accountability in ways that others put on “party lists” would not be. I don’t advocate any “party list” at all and do not believe political parties can be trusted to create them in any fashion that would not be manipulated internally by their cronies/insiders.

The party-proportional votes can be weighted to overcome other factors such as the fact that some districts get full members despite having very low population (the Yukon, Nunuvat, NWT, Labrador, all four PEI seats): these should probably not also get influence in the party-proportional round(s) unless they deliberately trade away some of their guaranteed whole seats.

Why would they ever choose such a tradeoff? Because it may be the best or only way to get a member of an aboriginal or other minority party elected.
If the MMP round is cognizant of where the votes come from and where the particular party member elected ran, tendencies of party proportional seats to override regional preferences can be minimized. If robust enough such formulas can even be used to permit district borders to match natural bioregional boundaries which never change at all – dealing with changes in population by adding fractions of members as opposed to changing borders to add full members – those fractions being accomodated in the final and party proportional round. That eliminates gerrymandering altogether, allows the borders to exactly match fixed municipal and bioregional organizing units, and permits wild or sparsely populated districts to be represented by single members while remaining quite large. As noted it would also allow some flexibility, each district could for instance choose to be represented more party-proportionally and less by direct election of specific candidates, within some global limit. So, federally, if Northern Ontario decided that it was better served with 1.0 party proportional members than with one full MP elected directly, they could make that choice say given 60% support in a referendum. It seems less likely however that these very electorally priveleged regions would make such choices, more likely Toronto would decide that it needed more influence in the party-proportional/MMP round and allocate say 4.5 seats to that round while keeping 22 directly elected MPs via STV – making it worthwhile for parties to focus national campaigns on Toronto and ensuring some Conservatives will also be elected.
(dealing with the worst problem of FPTP, discouraging political talent and competition).

A complete electoral system that includes all these features is available for review by anyone who will commit in writing not to advocate it until June 2009 and not to advocate it at all in preference to IRV or simple STV. Unlike Mr. Brekke I will not be muddying the waters by showing it off now.

Until May of this year our best bet for reform is to back BC-STV in BC and IRV federally. If either succeeds it sets the pattern for other reforms – and we can revisit the question of party proportional seats in the future.