Sunday, December 18, 2016

Canada wants moderate proportionality.

Many people have said Canada needs “a moderate proportional voting system.” Sure, Canadians are not extremists. But what does that mean?

In the 2015 election, the Liberals famously got 39.5% of the vote. (Actually, they got 39.8% of the five-party vote.) Under perfect province-wide proportionality, they would elect 40.5% of the MPs, but First Past The Post gifted them a “winner’s bonus” of 47 seats. In 2011 FPTP gifted Harper a “winner’s bonus” of 40 seats, in a smaller House.

On the votes cast in 2015, if you exclude Green votes outside BC where they got less than 5%, the Liberals got 40.7% of the votes, and under perfect province-wide proportionality they would elect 41.7% of the MPs. In a conventional MMP model with 14-MP regions and counting all Green votes, the Liberals elect 41.3% of the MPs.

In a more moderate 8-MP-region model with 38% top-up MPs, the Liberals elect 42.9% of the MPs, a bonus of 8 seats. Way better than the actual bonus of 47 seats. We can live with that.

The ranked ballot in single-member ridings is off the table.

One thing “moderately proportional” does NOT mean is the ranked ballot in single-member ridings. The multi-party Electoral Reform Committee’s majority report said the choice is between a good proportional system and First-Past-The-Post. The Liberal minority report did not even mention the ranked ballot. The Committee Chair, Liberal Francis Scarpallegia, said “no one wants the ranked ballot.” A huge step forward.

Justin Trudeau promised to make every vote count.

A recent Environics poll found 67% of Liberal voters feel the Liberal government should keep its promise and move forward with reforming Canada’s voting system. Only 10% disagreed, while 23% were unsure.

Canadians expect him to deliver this promise in full and on time.

A Scott Simms-inspired model?

In a recent discussion with Newfoundland MP Scott Simms, former Democratic Reform Critic for the Liberals, Simms agreed 10% top-up MPs was way too light, but suggested 20%. Lord Jenkins’ Report in the UK recommended 15% to 20% top-up MPs. I cannot imagine two people more different that Lord Jenkins and Scott Simms, yet they have the same thought.

For example, say we give each province 18.7% top-up MPs. Suppose we start by adding 44 top-up MPs to the House, by giving each province 10.5% more MPs and rounding the number up. The neat feature of these numbers is that they give the Atlantic provinces the 18.7% we want. The present 32 Atlantic ridings are unchanged, like the 3 ridings of the Territories. (The 44 new seats includes 3 more MPs for the Territories.) In the rest of Canada, we make enough of the present ridings bigger to give each province 18.7% top-up MPs. This adds another 28 regional MPs, cutting the number of local ridings to 310, each only 10% larger (outside Atlantic Canada). In total we have 72 regional top-up MPs.

I have done a simulation. With 32 regions, each with about 11 MPs today, outside Altantic Canada they will each generally become 10 local MPs and 2 regional top-up MPs.

Overly moderate, too moderate for me. And yet, no false majority on the votes cast in 2015. The Liberals get a bonus of 29 seats, but are 8 MPs short of a majority in the larger House. Interesting to look at.

This model still has many of the benefits of proportional representation. Liberal voters now unrepresented elect MPs in non-metropolitan Alberta, Vancouver Island, and the Barrie—Owen Sound region. And under-represented Liberal voters elect more MPs in Saskatchewan, Calgary, Edmonton, the BC Interior and North, and the London—Windsor region. Conservative voters unrepresented in five of Quebec’s seven regions elect MPs, as do those in Toronto, Peel Region, Northern Ontario, the north half of Metro Vancouver, Vancouver Island, and Yukon. The Atlantic Provinces have six opposition MPs. NDP voters everywhere outside PEI are represented. Even Green voters in Vancouver, Manitoba and west-central Ontario elect MPs, and 14 in total after the Green vote doubles under PR.

And as Prof. Dennis Pilon says in this video "Now keep in mind that, when you change the voting system, you also change the incentives that affect the kinds of decisions that voters might make. For instance, we know that, when every vote counts, voters won't have to worry about splitting the vote, or casting a strategic vote. Thus, we should expect that support for different parties might change."

“A moderate system”

Prof. Nathalie Des Rosiers told the Electoral Reform Committee about the Law Commission Report on Aug. 22 “We were trying to maintain the good parts of the first past the post system while remedying the bad parts. It was a moderate report that was aimed at helping Canadians and Parliament grapple with this issue of electoral reform.” She mentioned one-third top-up MPs.

Alex Boulerice responded “as the Scottish model shows, even in the Westminster tradition, changes can be made toward a moderate proportional voting system. I don't think anyone here would want our system to become extreme.

New Zealand’s 2012 MMP Review Commission said “The system of MMP adopted by New Zealand in 1993 is a moderate form of proportional representation which seeks to balance two important objectives. One is the principle of proportionality: that a party’s share of seats should reflect its share of the nationwide vote. The other is the need to ensure elections deliver effective Parliaments and stable governments by avoiding an undue proliferation of very small parties in Parliament.”

Oddly, the Electoral Reform Committee never heard one witness advocate the full Jenkins Commission Report, which was only 15% to 20% top-up MPs; deliberately very moderate. Jenkins wrote:
“In considering the level of Top-up we are required to balance carefully the potentially competing criteria set out in our terms of reference. On the one hand the importance of maintaining the link between MPs and their constituencies and the need to ensure stable government - to the arguable extent that this requires single party majority government most of the time - pushes towards keeping the level of Top-up as low as possible. On the other hand the requirement to deliver broad proportionality would push us towards a larger Top-up sufficient to correct, or at least substantially to ameliorate, potential disproportional outcomes on the constituency side. . . . a Top-up of between 15% and 20% of MPs would do sufficient justice to the three competing criteria discussed above to be acceptable.
. . . . . . . without producing any likelihood of a stagnant and unhealthy prospect of constant and unchangeable coalition.”

Electoral Reform Committee Chair Francis Scarpaleggia said on CBC Dec. 7: “I think you would want a moderate system of proportionality that would still allow for majority governments.” I expect he meant the same thing as Jenkins: a moderate level of proportionality that would still allow for some single-party majority governments.

So we can promote moderate proportionality, as long as it is fairly proportional. Many PR countries have a Gallagher index much less than five. Canada is a moderate country, after all.

But not too moderate

But this Scott Simms-inspired model is too moderate for me. Interesting to look at, though. 

Monday, December 12, 2016

Could Canada take an "incremental approach" to proportional representation?

The Electoral Reform Committee acknowledged that the overwhelming majority of testimony was in favour of proportional representation.

The choice, it found, is between keeping the present system or adopting a good proportional representation system that maintains the connection between voters and their MP. That would be the mixed-member proportional system described in detail here

An "incremental approach?"

Could an “incremental approach” to an MMP model work? This trial balloon was floated by the NDP and Green members of the Electoral Reform Committee.

This is not the best solution. It could mean waiting until 2025 for full implementation. But let's see what it would mean.

There will be a redistribution after the 2021 census. With the growing population of several provinces, I think there should be 35 more MPs from the growing provinces. Otherwise, the six smaller over-represented provinces will become even more over-represented.

An incremental approach is to keep the present ridings for 2019, and add 38 MPs to the House (including 3 extra MPs for the Territories) as top-up MPs for the 2019 election and the next election (2023 or earlier).

The second phase is to adopt, by legislation in 2017, a full mixed-member proportional representation system, to be implemented along with the recommendations of the Electoral Boundaries Commissions after the 2021 census, which are likely to take effect May 1, 2024. Again, it would have 376 MPs.

MMP-lite

With only 11% top-up MPs, the first phase is an “MMP-lite” model.

On the votes cast in 2015, the projected results would have been 189 Liberals, 112 Conservatives, 56 NDP, 13 Bloc, and 6 Greens. By contrast, a fully proportional model would have elected 153 Liberals, 119 Conservatives, 76 NDP, 17 Bloc, and 11 Greens.

With 376 MPs, that’s a Liberal majority of only 1, compared to the current majority of 15. To have a stable government, a coalition or accord would be advisable. Canada has seen ten coalition governments, and six stable Liberal minority governments. With an election in 2019 on this model, a false majority government would still be possible, but rather less likely.

Although this first phase model is poorly proportional, it still has some of the benefits of a proportional system.

Liberal voters are better represented in Alberta, Saskatchewan and Southwest Ontario, with five more MPs from those regions. Strong near-winner candidates in southwest Ontario like Kimberley Love and Allan Thompson or Katie Omstead or Stephen McCotter would be in Parliament. So would a strong candidate from the north half of Saskatchewan like Tracy Muggli or Lawrence Joseph. So would strong near-winner candidates from Alberta like Matt Grant and Karen Leibovici.

Green voters are better represented, with MPs like Gord Miller and Bruce Hyer from Ontario, Jo-Ann Roberts and Frances Litman or Ken Melamed from BC, and Daniel Green or JiCi Lauzon from Quebec.

Conservatives in Atlantic Canada elect three MPs instead of being silenced. Conservatives in Metropolitan Montreal and western Quebec elect four strong candidates like former MLA and mayor Robert Libman, Moroccan-born lawyer Valerie Assouline, business community leader Jimmy Yu, and Lebanese-Canadian architect Roland Dick.

New Democrats would have re-elected  Peggy Nash, Craig Scott and Andrew Cash in Toronto, Paul Dewar in Ottawa, Wayne Marston in Hamilton, and Jack Harris in Newfoundland, and added two new MPs in Ontario and two in Alberta.

Second phase

The second step is legislation in 2017 for a full mixed-member proportional representation system. It could have more than 33% regional MPs. The only reason to keep the number to 33% was to follow the present boundaries whenever possible, usually making three ridings into two. This is irrelevant with redistribution after the 2021 census, which are likely to take effect May 1, 2024.

The six small over-represented provinces would have temporarily gained a top-up MP, which ends with redistribution. Since they are over-represented, they are unlikely to complain. Using 2021 estimated populations from Statistics Canada, the other provinces will get: 83 for Quebec (close to the 86 they had temporarily), 138 for Ontario (up from the temporary 134), 49 in BC (up from 46), and 40 in Alberta (up from 38).

Technical note on projected size of the House:

The “electoral quotient” was set at 111,166 for the redistribution after the 2011 census, with a national population of 34,482,779. For the redistribution after the 2021 census the electoral quotient will be adjusted to reflect average provincial population growth since the previous redistribution. It might be increased to as much 123,812, but that would make Quebec over-represented.

It might be 119,180. In that case, the next House would have 351 MPs, an increase of only 3.85%. 
With some options, the six small over-represented provinces would temporarily gain a top-up MP, but lose that seat with redistribution. Since they are over-represented, they are unlikely to complain. 
Quebec, however, could find it humiliating to gain Top-Up MPs in phase 1, and lose them all again in phase 2. A risky tactic. 

Better to set the size of the next House high enough that Quebec gets to keep five of its eight temporary top-up” seats. If it becomes 376 MPs, the quotient is 109,970. 

Friday, December 9, 2016

The open-list Mixed Member Proportional system for which the Electoral Reform Committee found consensus.

The multi-party Electoral Reform Committee had a great deal to say about the Mixed Member Proportional System.

This could be the best model for Canadian democratic proportional representation.

The Committee found consensus on proportional representation

In Recommendations 1, 2 and 12, the Committee acknowledged that, of those who wanted change, the overwhelming majority of testimony was in favour of proportional representation. The Committee recommends that the Government should develop a new electoral system with a minimal level of distortion between the popular will of the electorate and the resultant seat allocations in Parliament, but not sever the connection between voters and their MP.

The Liberal minority report did not even mention the alternative of the ranked ballot in single-member ridings. At the press conference on the filing of the Committee Report, Liberal MP Francis Scarpaleggia (Chair of the Committee) confirmed "no one wants the ranked ballot." A huge step forward.

This is all consistent with the Mixed Member Proportional system, with open lists for the regional top-up MPs. You have two votes. Your first vote is for your Local MP. Your second vote is for a candidate for Regional MP. This counts as a vote for that candidate’s party. It helps elect region-wide MPs for top-up seats.

The Committee’s Report found consensus on open-list MMP.

Most individuals who favoured reform expressed support for this system (MMP).” “A majority of participants who advocated for electoral system change proposed the adoption of an MMP system, suggesting that it maximizes voter choice.” “Moving to an MMP system would keep the electoral system relatively simple. The local representation factor seems very familiar and similar to what [we] know with the current first-past-the-post system. It feels relatively simple and accessible on the ballot.” “Most respondents to the e-consultation strongly supported or supported the view that voters should determine which candidates get elected from a party’s list.”

The Report also discussed the details of MMP design: “In 2004, the Law Commission recommended two-thirds of MPs be elected in constituency races and the remaining one-third be elected from provincial or territorial party lists. The Commission noted that avoiding increasing the size of the House of Commons was a priority in determining said ratio.” “One way some countries with MMP systems have addressed the threat of the election of “fringe” or “extremist” parties is through the use of thresholds. For example, to be eligible to receive a share of the party vote seats in New Zealand, a party must garner at least 5% of the national vote.” Prof. Tanguay noted a built-in kind of threshold with MMP: “You'd need, probably, at least 10% of votes in a region to get one of those list seats.” As for the three Territories, “Some suggested adding a second compensatory MP to each territory to allow for some degree of proportionality” as the Law Commission recommended in 2004.

Further details are found in the Supplementary Report of the NDP and Green Party: “(MMP) with 2/3 of the House of Commons elected to represent direct constituencies, and 1/3 elected as regional compensatory members.” A group of three ridings will become two larger ridings each 50% bigger: “As such, since it would not affect current riding boundaries, a full riding redistribution would be unnecessary.”

The six smaller provinces have 60 MPs, an average of ten each. Local regions of about ten MPs match Prof. Tanguay’s comment above. The Hon. Stéphane Dion has advocated regions of similar size, to prevent creating different political micro-climates in different regions.

An MMP model with 34 local regions, with an average of 9.85 MPs each for the 335 MPs from the ten provinces, plus six MPs for the Territories, looks very practical.

Every region is represented

As Stéphane Dion likes to say “the whole spectrum of parties, from Greens to Conservatives, must embrace all the regions of Canada. In each region, they must covet and be able to obtain seats proportionate to their actual support. This is the main reason why I recommend replacing our voting system.” And Fair Vote Canada says “we must give rural and urban voters in every province, territory and regional community effective votes and fair representation in both government and opposition.”

In all 34 regions, on the votes cast in 2015, voters for all major parties have a representative. For example, Liberal voters who are not represented today in Vancouver Island, South and Central Alberta, Northern Alberta, Windsor—Sarnia, and Barrie—Owen Sound will elect MPs. So will Conservative voters in Atlantic Canada, Montreal and the western 72% of Quebec, Toronto, Peel Region, Northern Ontario, the north half of metropolitan Vancouver, and Vancouver Island.

How will regional MPs operate? 

Most regional MPs will each cover several ridings. Take Saskatchewan as an example. On the votes cast in October 2015, Liberal voters there would have elected two regional MPs. They might be based in Saskatoon and Prince Albert, but they would likely have additional offices in North Battleford and elsewhere, just as MP Kelly Block has offices in Martensville, Humboldt and Rosetown. This is just the way it’s done in Scotland, where each regional MP normally covers three local ridings, and holds office hours rotating across them. 

Even the Ministry of Democratic Institutions noted (no longer on-line) “Of the 34 member countries of the Organisation for Economic Co-operation and Development (OECD), Canada is one of only three that continue to use the FPTP system to elect legislators.” The rest mostly use proportional representation and have stable majority coalition governments like Germany.

Not my model

The model I am describing is not my personal model. I would use more than 33% regional MPs. This is the model I have taken from the reports outlined above.

Is one-third regional MPs enough?    

My simulation of this model on the votes cast in 2015 results in 145 Liberals (a bonus of 8 MPs over the perfectly proportional result), 107 Conservatives (short 3), 69 NDP, 15 Bloc, and 5 Greens (short 5). Not perfect, but reasonably proportional.

Prof. Byron Weber Becker has run the model on his software. It has a Gallagher Index of 2.94, and a Composite Gallagher Index of 4.25.

Looking at provinces, Ontario shows a Liberal bonus of 4, 2 from the Greens, 1 from the Conservatives, 1 from the NDP. This is because of the Liberal sweep of Toronto, Peel Region and Oakville, where 52% of the vote gave them all 37 MPs. With only 33% compensatory MPs, the Liberals get a bonus of 4 MPs there, 3 from the Conservatives, 1 from the NDP. But another region gives the Conservatives a bonus, and the other eight regions of Ontario show fully proportional results.

New Brunswick shows a pattern similar to Toronto: a sweep on 51.6% of the vote, resulting in a Liberal bonus of 1 MP from the Conservatives. Conversely, the NDP sweep of six of Vancouver Island’s seven seats means BC shows an NDP bonus of 1 from the Conservatives.

Quebec is close to perfect: a Liberal bonus of 2 and a Conservative bonus of 1, 2 from the Greens, 1 from the NDP. In Alberta’s four regions, rounding anomalies give the Liberals and NDP a bonus of 1 each, 1 from the Conservatives, 1 from the Greens.

What if the Green vote doubles?

Another way to test whether 33% regional MPs is enough, is to project the outcome if the Green Party vote doubles, as they expect it would under PR.

A projection showing their vote doubled from non-voters (everywhere but in Elizabeth May’s riding of Saanich—Gulf Islands) shows them electing the correct 22 MPs. They elect MPs in every province but Newfoundland & Labrador and P.E.I. Again, the sweeps in Toronto and New Brunswick give the Liberals a bonus of 11 MPs, of which 7 are from the NDP, 2 from the Bloc, and 2 from the Conservatives. Not perfect, but reasonably proportional.

Which ridings would change?

Every local region, with about 10 MPs per region, will still have the same number of MPs as it does today. Those MPs will become 67% local, 33% regional. Wherever possible, three present ridings will become two larger ridings 50% larger. In my simulation, I use 14 nine-MP regions, 6 twelve-MP regions, 4 six-MP regions, and one 15-MP region. The Boundaries Commissions will make short work of this. Exceptions will be in only about 14% of ridings, about 38 of the present 275 in the larger provinces. Seven ridings could be “grandfathered:” the three for the Territories, and four more remote and aboriginal ridings.

Law Commission of Canada

In 2004 the Minister of Justice, the Hon. Irwin Cotler, tabled the Law Commission of Canada’s Report recommending a mixed member proportional system, just as outlined above. MMP is used in Germany, Scotland, New Zealand, and other jurisdictions. 

The Law Commission recommended one vital improvement: no closed lists. All MPs are elected and have faced the voters. If voters for a party are entitled to elect a regional MP, it will be the party’s regional candidate who got the most votes across the region. 

You have two votes. With your first vote, you help elect a local MP as we do today. With the second, you also help elect a few regional MPs: it’s proportional. You can cast a personal vote for the regional candidate you prefer: it’s personal. Using a ballot like the one that won the PEI referendum in 2016.

The regional candidates will be democratically nominated in 34 local regions, each small enough to make them accountable, not by the province-wide lists that Ontario voters rejected in their 2007 referendum.

Ranked ballots?

Could this model use a ranked ballot to elect the local MPs? In most ridings this would make no difference, but it would increase the Liberal bonus to the extent that the Gallagher Index would be 5.93, higher than the target of 5, with a Composite Gallagher Index 6.82. Not recommended.

Technical note

The calculation for any PR system has to choose a rounding method, to round fractions up and down. I have used the “largest remainder” calculation, which Germany used until recently, because it is the simplest and most transparent. In a 10-MP region, if Party A deserves 3.2 MPs, Party B deserves 3.1, Party C deserves 2.3, and Party D deserves 1.4, which party gets the tenth seat? Party D has a remainder of 0.4, the largest remainder. In a region where one party wins a bonus (“overhang”), I allocate the remaining seats among the remaining parties by the same calculation.    

Appendix - the regions

My simulation using this model lets voters elect 220 local MPs and 115 regional MPs in 34 regions, plus two each from the three Territories.

The regions are:

Ontario:
Toronto Central—Scarborough 13, North York—Etobicoke 12, York—Durham 15, Mississauga—Brampton—Oakville 12, Hamilton—Niagara—Halton 12, Central Ontario (Barrie—Owen Sound) 8, Waterloo—Brantford—Wellington—Dufferin 9, London—Oxford—Norfolk 6, Windsor—Sarnia 6, Central East (Peterborough—Kingston) 9, Eastern Ontario 10, Northern Ontario 9.

Quebec:
Montréal-est 9, Montréal-ouest 9, Laval—Lanaudière 9, Laurentides—Outaouais—Abitibi-Témiscamingue—Nord 11, Rive-sud 9, Montérégie-est—Bécancour—Estrie 9, Québec City—Mauricie—Saguenay-Lac-Saint-Jean—Côte-Nord 13, Chaudière-Appalaches—Bas-Saint-Laurent—Gaspésie 9.

Alberta:
Calgary 10, South-Central Alberta 9, Edmonton 9, North Alberta 6.

British Columbia:
Vancouver—Burnaby—Coquitlam—Maple Ridge 12, Surrey—Fraser Valley—Richmond—Delta 12, Vancouver Island, North and West Vancouver 9, BC Interior and North 9

Manitoba, Saskatchewan, New Brunswick, Nova Scotia, Newfoundland & Labrador, and Prince Edward Island: each are one region.

2019 simulated result

As I said, my simulation of this model on the votes cast in 2015 resulted in 145 Liberals (a bonus of 8 MPs over the perfectly proportional result), 107 Conservatives (short 3), 69 NDP, 15 Bloc, and 5 Greens (short 5). Not perfect, but reasonably proportional.

The simulated result of the 2019 election was similarly close to proportional: 125 Liberals (a bonus of 8), 118 Conservatives (short 1), 53 NDP (short 3), 26 Bloc, 18 Green (short 4), and  one Independent. This is because one-third regional MPs is not quite enough, but with 341 MPs, a Liberal-NDP coalition government would have a majority. Both parties would have seats in every province, except no NDP MP from PEI. And the Greens would have MPs from every province except Alberta, Saskatchewan, and Newfoundland & Labrador.

(This post revised Sept. 16, 2020.)