Tuesday, March 16, 2010

The Ontario mixed member model the Citizens’ Assembly almost chose.

When Fair Vote Canada members first met Kingston’s cabinet minister John Gerretsen back in 2004, we didn’t have to explain the Mixed Member Proportional (MMP) system to him. He explained “the German system,” as he called it, to us. We later found he had been pushing for it since he was first elected in 1995 (and even since 1986 when he was President of the Association of Municipalities of Ontario). When the Liberals finally won in 2003, they had spent 60 years in the political wilderness minus only the five years from 1985-90 -- and in 47 of those years they were facing a government with a fake majority supported by a minority of voters. They remembered for a few years why they needed PR.

In “the German system” you have two votes, and more choice. We still elect majority of MPPs locally. Voters unrepresented by the local results top them up by electing regional MPPs. The total MPPs match the vote share. With the regional "Open list" version, voters can vote for whomever they like out of the regional candidates nominated by the party's regional nomination process. Like the right-hand part of this ballot.

See MMP Made Easy.

To compensate for the disproportional local results we know all too well, the party’s voters elect personally some regional MPPs. They are the party's regional candidates who get the highest vote on the regional ballot. So the voter casts one vote for local MPP, and one for their party and (if they wish) for their favourite of their party's regional candidates. An exciting prospect: new voices from new forces in the legislature, and the voters have new power to elect who they like.

John Gerretsen was quite specific. The top-up MPPs should be elected regionally, and the regions should not be too large. Kingston should not be lumped in with Ottawa, he said. Those who know Eastern Ontario know that the mid-eastern and Lake Ontario regions and the Ottawa region have many divergent interests, so we were not surprised when Gerretsen mentioned one or two of them.

However, the model put to voters in the October 2007 referendum, designed by the 103 members of the Ontario Citizens Assembly (CA), had province-wide closed lists, not the mid-sized regions John Gerretsen had told us he wanted, and that Ontario NDP policy had favoured since 2002. Ontario voters rejected closed province-wide lists in the 2007 referendum.

What would the 2011 election results would have been under the model John Gerretsen wanted? And the 2014 results?

Not enough time

The original schedule prepared by the Democratic Renewal Secretariat called for the process to begin one year earlier. The legislation was to be passed in the spring of 2005. Due to an end-of-session legislative logjam, the PCs were able to demand a Select Committee be inserted in the process. Although it produced an excellent report, the resulting delay was fatal.

In May 2008 the CA’s Chair, George Thomson, spoke to the Annual General Meeting of Fair Vote Canada. He said that, if those 103 Citizens had had another six or eight weeks to deliberate, he felt some elements might have been different, like regional lists and open lists. (He thought the basic model would have stayed the same: 129 MPPs, 90 local, 39 top-up.) Those additional weekends were not possible due to the tight schedule.

Those candidates on province-wide lists were to be nominated democratically by parties, but in the few months between May and the referendum, no major party -- not even the NDP -- had enough time to design a nomination system. The model’s opponents -- even, ironically, an appointed Senator -- said it sounded like parties would appoint those 39 MPPs.

The public had not enough time to understand the CA’s recommendation, even if it had been publicized properly. Adding injury to injury, the Ontario government surprised the Citizens Assembly by refusing to distribute their Report, unlike the BC process which they had otherwise been following. Clearly the majority of cabinet had lost interest in their democratic reform effort.

Nine mid-sized regions

What would their model have looked like, with those mid-sized regions?

The North would have been a separate region. It could have had a special feature: it could have kept unchanged the ten present ridings north of the French River, and added only two regional MPPs.

The City of Toronto could have gone from 22 local MPPs to 25, 17 local and 8 regional.

The other seven regions would have had 12 to 14 MPPs each, such as 9 local, 4 regional.

How would regional MPPs serve constituents? Here's how it works in Scotland.

Open list

Since local candidates can also be on the regional half of the ballot, voters might have had ten or so of their party’s regional candidates to choose from, but not the “bed-sheet ballot“ found in some countries. So voters would have a real choice among a manageable number of competing candidates from the party they support. And they could also choose to vote just for their party, leaving the candidates ranked as their party’s nomination process had done. That's the variation of "open-list" recommended by the Law Commission of Canada, known as "flexible list."

The flexible open list method was also recommended by the Jenkins Commission in the UK. Their colourful explanation accurately predicted why closed lists would be rejected in Canada: additional members locally anchored are “more easily assimilable into the political culture and indeed the Parliamentary system than would be a flock of unattached birds clouding the sky and wheeling under central party directions.”

The most recent official Quebec study on the topic also looked favourably at regional open list MMP.

More women and minorities

With a choice of your party’s candidates on the regional ballot, we would elect more women. Polls show 94% of women voters want to see more women elected, but so do 86% of male voters.

And when parties nominate a group of candidates, not just one, they nominate more women. What regional convention, nominating five candidates, would nominate only one woman, or no minorities?

90 local ridings

Local ridings would be slightly bigger than today, but not so you’d notice. Often, ten present ridings would become nine. (But the North could have kept unchanged the ten present ridings.)

Closed province-wide lists?

So why did those 103 Citizens choose province-wide closed lists?

George Thomson’s comments in May 2008 show the process the 103 Citizens went through. Their big design problem was Ontario’s geography, and the fact that our local ridings are already too large. Until Mike Harris shrank the House in 1999 we had 130 MPPs, compared with 101 MPs at that time. Many members of the CA wanted to keep the present 107 ridings and add at least 36 “top-up.” Others wanted to keep 107 MPPs but have only 80 larger local ridings and 27 top-up. Others wanted a higher ratio of top-up. Their big achievement was consensus on 90 plus 39.

They had decided on province-wide lists early in the process, before they agreed on the numbers. Back at that point, many members wanted to use all the most proportional options in order to leave them free to have less proportional numbers of MPPs. For example, on those 2007 votes, because the Citizens’ model had only 30% “top-up” MPPs, it would have resulted in more than 55 Liberal MPPs. And then, making the lists regional rather than provincial added a further four more Liberal MPPs.

Still, once they had 39 top-up MPPs, regional lists became possible, and open list became possible. Four Liberal MPPs too many, in our 2007 example, would have been a modest price to pay for a more accountable and democratic model. But by the time they made that decision for 39 top-up MPPs, it was too late to go back and redesign.

This is no one’s fault. The Democratic Renewal Secretariat had planned for the whole process to start a year earlier. The legislature’s Select Committee got inserted into the process, and did a wonderful job, but that left both the CA and the public debate short of vital time.

Regional candidates

Why do I say voters would have at least five of their party’s regional candidates to choose from, maybe ten or so, when most regions elect only four regional MPPs?

Take a region with 13 MPPs, nine local, four regional. Suppose Party C’s voters cast 30% of the votes in the region, but elect no local MPPs, and suppose no other party’s voters earn a regional MPP. Party C’s voters elect all four regional MPPs. But if one of them dies or resigns during the legislature term, the regional candidate with the next highest votes moves into that seat. A party must run at least five, to have a spare.

This matters to women and minorities. A regional convention, nominating five candidates, would almost certainly nominate at least two women, and at least one cultural minority member.

On the other hand, suppose Party A’s voters cast 61% of the votes in the region, but elect only seven of the nine local MPPs. They also elect one regional MPP. But if the seven local winners were also on the regional ballot, the party needed at least nine regional candidates, one elected, and again one spare. To get good balance I can see them nominating ten regional candidates.

What would the legislature have looked like in 2007?

For another example, let’s see what the Ontario legislature would have looked like under this model if voters voted as they did in 2007.

This projection assumes voters voted as they did in 2007. In fact, if voters knew every vote would count, more would have voted -- typically at least 6% more. And some would have voted differently -- no more strategic voting. We would likely have had different candidates -- more women, and more diversity of all kinds. Who knows who might have won real democratic elections?

But on the votes as cast, we would have seen 61 Liberals, 39 PCs, 20 New Democrats, and 9 Greens.

Since the Liberals got only 42% of the vote, they would not have had an outright majority of the 129 seats. But as John Gerretsen said “Nobody is ever 100-per-cent right and nobody is ever 100-per-cent wrong. Governing is the art of compromise. There’s nothing wrong with having the governing party take into account smaller parties.”

Because the Citizens’ model had only 30% “top-up” MPPs, this 2007 projection is not perfectly proportional. That would have meant 55 Liberals, 41 PCs, 22 NDP and 11 Green. But it’s close enough: the potential coalitions are the same either way.

Progressive Conservative voters would have elected 13 more MPPs.

Toronto PC voters would have elected five MPPs, not none. No doubt leader John Tory and councillor David Shiner, and maybe school trustee Angela Kennedy or Bernie Tanz or Pamela Taylor or Igor Toutchinski or Lillyann Goldstein or Andy Pringle or Gary Grant?

Peel-Oakville PC voters would have elected three MPPs, not just one. Maybe Rick Byers and Pam Hundal or Tim Peterson or Nina Tangri?

Ottawa Valley PC voters would have elected five MPPs, not just three. Maybe Chris Savard from Cornwall, and Graham Fox or Trina Morissette. (However, Central East Region PC voters would have elected one fewer MPP.)

Hamilton area PC voters would have elected five MPPs from Hamilton, Niagara, Brant and Burlington, not just three. Maybe Chris Corrigan and Bart Maves or Tara Crugnale.

Southwestern PC voters would have elected three MPPs, not just two from the London-Windsor area. Maybe Monte McNaughton or Allison Graham.

Northern PC voters would have elected two MPPs, not none. Maybe Bill Vrebosch and Rebecca Johnson or Ron Swain.

NDP voters would have elected nine more MPPs.

Central East region NDP voters would have elected two MPPs, not none. Maybe Kingston’s Rick Downes and Belleville’s Jodie Jenkins or Peterborough’s Dave Nickle or Muskoka's Sara Hall.

Southwestern NDP voters would have elected two MPPs, not none. Maybe London’s Stephen Maynard and Sarnia’s Barb Millitt or Windsor’s Mariano Klimowicz?

Central West NDP voters would have elected one MPP from the area from Waterloo to Owen Sound, not none. Maybe Rick Moffitt or Catherine Fife or Paul Klopp?

Peel-Oakville NDP voters would have elected one MPP, not none. Maybe Glenn Crowe or Shaila Kibria or Mani Singh or Gail McCabe?

Ottawa Valley NDP voters would have elected one MPP, not none. Maybe Will Murray or Edelweiss D'Andrea?

York-Durham region NDP voters would have elected two MPPs, not none. Maybe Oshawa’s Sid Ryan and York Region's Nancy Morrison or Rick Morelli?

Toronto NDP voters would have elected five MPPs, not just four. Maybe Paul Ferreira or Sheila White or Peter Ferreira or Sandra Gonzalez?

Green Party voters would have elected nine MPPs.

Toronto Green voters would have elected two MPPs: maybe leader Frank de Jong, and Caroline Law or Dan King?

York-Durham Green voters would have elected one: maybe June Davies of Uxbridge or Liz Couture of Richmond Hill?

Central East Region Green voters would have elected one: maybe Judy Smith Torrie of Northumberland, or Matt Richter of Muskoka, or Simcoe's Peter Ellis or Erich Jacoby-Hawkins, or Kingston's Bridget Doherty?

Peel-Oakville Green voters would have elected one: maybe Deputy Leader Dr. Sanjeev Goel, or Rob Strang or Marion Frances Schaffer or Paul Simas?

Central West Green voters would have elected one: maybe Shane Jolley of Owen Sound or Ben Polley of Guelph or Victoria Serda from Huron-Bruce or Judy Greenwood-Speers of Kitchener-Waterloo?

Ottawa Region Green voters would have elected one: maybe Greg Laxton or Elaine Kennedy?

Hamilton area Green voters would have elected one: maybe Melanie Mullen from Niagara or Peter Ormond from Hamilton or Ted Shelegy from Brant.

Southwest Region Green voters would have elected one: maybe Brett McKenzie or Jessica Fracassi?


Raymond said...

Just a few comments in support of this model:

Mid-sized regions are used by all the main political parties in Ontario. There are between 6 and 9 regional organizing bodies within the Liberal, PC, NDP & Green parties.

With the next redistricting, it looks like there will be about 127 seats in the legislature ... so the Citizens Assembly wasn't being unreasonable in recommending 129 seats for Ontario.

The newly legislated redistricting commission would have the ability to recommend extra representation for northern Ontario, to accomodate extreme geography as required. (Parry Sound district should probably go with the northern region, because of common interests.)

Some asymmetry in the size & population of regions, as well as in the percentage of "top-up" seats per region, helps to get consensus.

A good choice of regional boundaries is important because "geographical affiliation" is connected with one's personal identity.

Wilf Day said...

Indeed, Raymond, the Ontario PC Party uses nine regions too.

The regional boundaries would be subject to adjustment by the Boundaries Commission, but I wouldn't leave it up to the Boundaries Commission to decide whether Northern Ontario is a separate region. Clearly it is.

If we wanted to grandfather the existing Northern ridings, I would not extend that to the Parry Sound -- Muskoka riding which is partly in the South, and Muskoka is oriented to Simcoe County. But that's a detail. The point is, that the regions be a manageable size so that "open list" doesn't give undue advantage to political "rock stars." And the northern ridings need not be any larger than today.

Jonathan said...

Hi Wilf
Interesting model and I agree that the Ontario Citizens' Assembly might have chosen a different model if it had more time -- but it may not have.

The assumptions that your well developed model make are that a) regional representation is important and that b) it's worth having regional representation at the cost of proportionality and c) that this model values would be better than a more "simple & practical" province wide tier. I do recall that some members thought that it would be difficult to explain the counting vs. allocation rules around regional MMP and that regional counting and regional allocation might not be very proportional in districts that have few seats.

It's important to remember that indeed, regional list MMP was considered but the members chose province wide for the reasons mentioned above among others.

It's also true that time was a constraint and that ratio between the tiers was an important design decision.

Having said that, your model certainly provides a strong example of how regional MMP might work in Ontario and I think most importantly, furthers the discussion of the trade-off between values that a system ought to have and their design elements.

Wilf Day said...

Jonathan, yes, there are trade-offs in a regional model.

The 2007 votes were rather extreme: I believe the Liberals would have elected 57 local MPPs under the CA's actual model, with an overhang of two over the 55 MPPs that a fully proportional model would have generated. Your staff said an overhang of one or two was possible but not too likely, and I agree: 2007 was a sweep to an unlikely degree.

Using nine regions -- no region has fewer than 12 MPPs -- gives the Liberals one more MPP by their near-sweep of Toronto as a separate region, and one more similarly in each of Peel-Oakville, Southwest and the North. By making special provision for the North to have only two regional MPPs, so as to maintain the existing local northern seats, I add still one more local Liberal MPP. So the regional structure brings the Liberals up to 62, rather than 57 or 58. A small deviation from proportionality that, in this instance, does not affect the shape of the legislature, as a trade-off for much more voter choice, no "appointed" MPPs, and full regional representation in each party caucus. A very good trade-off, in my view, and more saleable.

My model uses the same simple and practical counting rules as the province-wide model, as to determining the seats won by each party in each region. Provincial calculation with regional allocation, a difficult calculation, would have given the Liberals 57 or 58 MPPs, at the cost of lack of transparency.

Yes, the 103 Citizens decided against regional lists, on the assumption they could revisit that decision later. No "decision tree" is perfect; decisions have to start somewhere. But in fact, once they decided on 39 top-up MPPs, they did NOT have time to revisit regional lists, or open lists. If the process had started a year earlier, they would have.

Raymond said...

Jonathan makes a good point about having clear values when designing an electoral system.

I think it may have been useful for the Ontario Citizens Assembly to hear from the political parties on the topic of regional representation. Specifically, each party could explain why is important to them, and why they organize their parties the way they do.

I got the impression from reading northern media commentary on the referendum that the regional "weight" of representation in the legislature was the biggest issue against MMP, and not necessarily the size of northern ridings. If true, that could leave some room to expand the size of local ridings a bit more, in favour of the added proportional tier.

That way, both proportionality AND regional representation would both be given significant importance.

Also, a 6-region model would generally be slghtly more proportional than a 9-region model, but there would be trade-offs.

Accountability could be increased by legislating "open" regional nominating conventions to "order" the party lists, as they do in Germany.

Wilf Day said...

Yes, Raymond, northern comment was against province-wide lists. They wanted to be sure the North wouldn't lose weight.

But it was also against larger riding sizes. This was a big issue in the whole North-East ever since 2002 when the federal boundaries commission proposed removing one northern seat. The resulting Algoma--Manitoulin--Kapuskasing riding was obviously absurd. The northern backlash was such that all parties promised in the 2003 provincial election to maintain the existing ridings (set in 1996) for provincial elections, and the government kept that promise. Every northern politician or activist I spoke to said maintaining the existing ridings was non-negotiable.

A 6-region model, with only four 23-MPP regions in southern Ontario outside the city of Toronto, would be pretty brutal. A 7-region model, with only five in southern Ontario outside Toronto, would be possible. Kingston would be with Ottawa, Kitchener-Waterloo would be with Hamilton, and I've been told that neither of those is a really good idea, but it would be workable. But it actually wouldn't change the problem of the Liberals sweeping Toronto, the Southwest and the North. That's why I think nine regions is best, as long as every region (except the North) has at least four regional MPPs.

As for legislating democratic regional (and local) nominations, as they do in Germany, it was outside the CA's mandate. But I'm all for it. See my blog post:

Raymond said...

I would agree that the smaller the regions (or ridings) work out more neatly in boundary designs. So I guess the issue is where to draw the line and why.

The most important design issue is to solve the two political "hot potatoes" of northern Ontario and Toronto by giving them separate regions.

Could a government give the "boundaries commission" a directive to observe the "recommended" regional boundaries, or would that be totally up to the commission?

Wilf Day said...

Raymond, I see Scotland and Wales reviewed their ridings and regions in 2007-8. Their Commission had to keep 73 local ridings in Scotland, 40 in Wales, and equal-sized regions: eight in Scotland with seven regional MSPs from each, and five in Wales with four regional AMs from each. The Boundary Commissions only had to group the ridings into regions, not too difficult.

In Ontario I would expect the rules would tell the Boundaries Commission how many local ridings to create, how many regional MPPs, that the North is a region, that the City of Toronto is a region, and that they are to design seven more regions with, if possible, four regional MPPs from each region. Of course, in designing both ridings and regions, community of interest is the prime consideration. I would also suggest a 10% limit on deviation from average riding populations, except in exceptional circumstances.

Arita said...

Hi there Wilf,
good to see there's still lots of interest out there.
During our our learning phase I spent quite some time on the "regions" I came to the conclusion that we(I) should have selected regional lists with at least 7 regions. But as the group decided on province wide lists thats what we went with. I do agree regarding the regions that it might have been a different outcome if we had more time.

I agree with Raymond that the boundaries should take into account the "geographical affiliation"

I also agree with you that the north is definately a region unto itself, and as such requires separate attention.

There are so many woulda coulda shouldas out there. In the end I don't think Ontario was ready for a change.
Many, many people in my riding stil vote the same as their grandparents did because thats the right thing to do. Grandpa would roll over in his grave if he knew you were voting against his party.

Keep up the good fight. We need a change from what we have and the sooner the better.

Wilf Day said...

Yes, Arita, many voters vote by tradition. But in 2007 Ontario had 8,557,653 people on the voters list. Of those, 21.8% voted Liberal, 16.3% voted Progressive Conservative, 8.7% voted New Democrat, 4.1% Green, and 47.9% did not vote. That’s a lot of non-traditional voters. Many of them correctly decided their vote would not count.

We didn’t reach them in the referendum campaign. My favourite anecdote was a hairdresser who always debates politics with a friend of mine. My friend was getting her hair cut the day after election day. “What was that second ballot for?” the hairdresser asked my friend. “The one about Hydro.” Hydro? A poll officer had told her “electoral reform” which she misheard as “electrical reform.”

Wilf Day said...

In the most recent election for the Scottish Parliament, using 8-region MMP, the Scottish National Party won a majority of seats on a minority of votes. Was this because of using eight regions?

Not really. Using eight regions gave the SNP one overhang seat. The real problem was that, unlike the Ontario Citizens' Assembly model, Scotland uses the "highest average" calculation, which Germany abandoned decades ago; it favours large parties. The other methods are "highest remainder" which has the virtue of simplicity, and "Sainte-Lague" which is the best mathematically. A new study has just shown that, if Scotland changed to Sainte-Lague but kept its eight regions, that would generate the fairest result, and the SNP would not have won a manufactured majority.