Sunday, April 25, 2010

MMP Made Easy

Scottish, Welsh, German and New Zealand Parliamentary elections use a type of Proportional Representation called the Mixed Member Proportional system (MMP).

The ten-second definition of MMP is this:

We still elect local MPs. Voters unrepresented by the local results top them up by electing regional MPs. The total MPs match the vote share.

MMP is a voting system which mixes our winner-take-all system with an element of proportional representation, so that the number of MPs elected to Parliament from each province matches the share of the overall votes cast by supporters of each party in that province.

Different places use different MMP models. This is a description of an MMP model with “open lists.”

Each voter has two votes.

The local vote is used to elect an MP to represent your riding, as today.

The regional vote or party vote is used to elect several regional MPs from your region.

The local vote can be cast by marking your ballot with an X for any candidate standing in your riding, as we do today. The candidate chosen by the largest number of voters in a riding wins the seat on a winner-take-all basis.

The regional vote can be cast by marking your ballot with an X for any regional candidate standing on the regional ballot.

If that candidate is a party candidate, this vote counts as a vote for your party. The parties' regional votes are then counted to give the level of support for each party in the region.

If a party’s voters have managed to elect only a few local MPs in that region, or none at all, that party gets additional “top-up” seats to make their final total more in line with their vote share in the "top-up" region.

The party’s regional candidates with the most votes win those seats. That’s why it’s called “open list.”

Every voter has competing MPs: you can go to your local MP or one of your diverse regional MPs. Germans call this "personalized proportional representation."

Your local ballot will look like today's ballot. Your second ballot -- the regional ballot --will look like the right-hand part of this ballot.

Here are further details on this model as designed by the Law Commission of Canada.

What would the 2011 election results have been under this model?

(The number of regional MPs you have depends on the size of your "top-up" region: if they are medium-sized regions it might be five, maybe as many as eight or more. In smaller provinces, the "top-up region" would be the whole province.)

Saturday, April 24, 2010

What would the British Columbia legislature look like under a fair voting system?

What would the British Columbia legislature look like under a fair voting system?

There are two likely options for a fair voting system for BC provincial elections. The British Columbia Citizens’ Assembly on Electoral Reform designed two systems in 2004.

BC-STV

They decided BC-STV was the best for BC, so that was the system that went before voters in a referendum, in 2005 and again in 2009. In 2005 voters were asked “Should British Columbia change to the BC-STV electoral system as recommended by the Citizens' Assembly on Electoral Reform?” They voted 58% yes, but the government had set a 60% threshold for success. In 2009 voters were asked “Which electoral system should British Columbia use to elect members to the provincial Legislative Assembly? ▪ The existing electoral system (First-Past-the-Post) ▪ The single transferable vote electoral system (BC-STV) proposed by the Citizens’ Assembly on Electoral Reform." This time they voted only 39% for BC-STV.

However, a poll after the referendum showed that 44.3% of those who voted for first-past-the-post in the referendum responded they are in “favour of replacing first-past-the-post with a voting system in which the percentage of seats a party gets in the legislature is more in line with their percentage of the popular vote.” That makes 66% of BC voters in favour of some proportional system.

Why not STV?

STV has been used in Ireland since 1922. It’s the only thing Northern Ireland and the Irish Republic agree on; Northern Ireland has used it since 1973. And it has now spread across the sea to be used in Scotland’s local elections. To those familiar with STV, it’s an excellent system if the district magnitudes are large enough for decent proportionality, like Northern Ireland’s six-MLA districts, and the population per district is low enough, like Northern Ireland’s 98,000 people per district. Unfortunately BC voters are not familiar with STV. If the British Liberal Democrats had succeeded in moving the UK towards STV, this might have changed; but even they gave up.

Furthermore, the BC-STV model designed by the Electoral Boundaries Commission was not very proportional. With each district having, on average, only 4.25 MLAs, Green Party voters, who deserved to elect seven MLAs on the votes cast in 2009, would have been lucky to elect even three MLAs. Yet many BC voters complained that the proposed 20 districts, with an average population of about 222,000, were already unmanageably-large. This is not a trivial point. In a country with as much geography as Canada, fitting our geography into the voting system is the major design issue.

Open-list MMP

The Citizens’ Assembly (CA) designed, before they chose STV instead, an excellent Mixed Member Proportional (MMP) system. The majority (60%) of MLAs are elected in local districts like today’s. The others are "top-up" regional MLAs: to compensate for the disproportional local results we know all too well, the party’s voters elect personally some regional MLAs.

See MMP Made Easy.

This is the system invented by British political scientists in 1946 in the British Zone of West Germany. It took the old German proportional representation system and grafted British personal MPs into it. “Personalized proportional representation” the Germans called it. “The best of both worlds” said political scientists.

With the CA’s regional "Open list" version, voters can vote for whomever they like out of the regional candidates nominated by the party's regional nomination process. The elected Regional MLAs are the party's regional candidates who get the highest vote on the regional ballot. You have two votes: one for local MLA, one for regional MLA which counts as a vote for your party. The German province of Bavaria does this too.

If the CA had chosen the “flexible list” variant, where you can vote for the list or for an individual on it, the ballot would have looked like the one recommended by the British Independent Commission on the Voting System (the Jenkins Commission). The voter casts one vote for local MLA, and one for their party and (if they wish) for their favourite of their party's regional candidates. This same model was recommended for Scotland by the Arbuthnott Commission as an improvement on their MMP system; but no action yet. The result is much the same with any open-list model: all MLAs have faced the voters, and no one has a safe seat. (But I'd bet the CA would have chosen straight open-list, the Bavarian model.)

Power to the voters

An exciting prospect: voters have new power to elect who they like. New voices from new forces in the legislature. No party rolls the dice and wins an artificial majority. Cooperation will have a higher value than vitriolic rhetoric. One-party dominance by the Premier’s office will, at last, be out of fashion. Governments will have to listen to MLAs, and MLAs will have to really listen to the people. MLAs can begin to act as the public servants they are.

Based on the Ontario and BC experience, many reformers now think open-list MMP with regional lists is the only system likely to be acceptable to Canadians.

Local districts and regions

The 51 local BC districts would each have about 87,000 people (smaller in the North, no doubt). The CA ran out of time before settling details like the number of regions, which might have been four, five or six; I’m using six, electing a total of 34 regional MLAs.

What would the legislature look like?

For an example, let’s see what the BC legislature would have looked like under this model if voters voted as they did in 2009.

This projection assumes voters voted as they did in 2009. 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 seen different candidates -- more women, and more diversity of all kinds. We could have seen different parties. Who knows who might have won real democratic elections?

But on the votes as cast in 2009, the overall result is 41 Liberals, 37 New Democrats, 6 Greens, and independent Vicky Huntington.

Competing MLAs

Instead of having only a local MLA -- whom you quite likely didn’t vote for -- you can also go to one of your regional MLAs. On this projection, all six regions will have at least one regional MLA from each of the three parties. Even Northern voters, assuming they elected three local Liberals and two local New Democrats, would have elected one regional MLA from each party. Even Interior voters, where I expect Liberal voters would have elected seven of the ten local MLAs, would also have elected one regional Liberal MLA. Even Vancouver Island voters, where I expect NDP voters would have elected seven of the nine local MLAs, would also have elected one regional NDP MLA. That’s because the CA wisely chose a model with 40% regional MLAs.

NDP:

Many NDP voters are under-represented. Voters in the Interior would have elected 16 MLAs (10 local, 6 regional), including two more NDP MLAs: maybe Doug Brown and Tish Lakes or Charlie Wyse or new candidate Lakhvinder Jhaj or Troy Sebastian?

Surrey-Delta-Langley-Fraser Valley voters would elect 17 MLAs (10 local, 7 regional), including two more NDP MLAs: maybe new candidates Debbie Lawrance and Pat Zanon or Bonnie Rai or Gwen O’Mahony or Lynn Perrin?

Vancouver-Richmond voters would elect 14 MLAs (8 local, 6 regional), including another NDP MLA: maybe Gabriel Yiu or Jenn McGinn or Helesia Luke?

Liberals:

Liberal voters on Vancouver Island are under-represented. Vancouver Island voters would elect 15 MLAs (9 local, 6 regional), including two more Liberal MLAs: maybe new candidates Marion Wright and Dawn Miller or Dianne St. Jacques or Cathy Basskin?

Green Party voters would have elected six MLAs.

Green Party voters in Vancouver Island would have elected a regional MLA, no doubt the leader Jane Sterk.

Vancouver-Richmond Green Party voters would have elected a regional MLA: maybe Damian Kettlewell or Vanessa Violini or John Boychuk or Jodie Emery?

Burnaby-Tri-Cities-North Shore-Maple Ridge voters would have elected 15 MLAs (9 local, 6 regional) including a Green Party regional MLA: maybe young Michelle Corcos or Helen Chang or Jim Stephenson?

Voters in the North would have elected 8 MLAs (5 local, 3 regional), including a Green Party regional MLA: maybe Liz Logan or Lisa Girbav?

In the Interior, Green Party voters would have elected a regional MLA: maybe Julius Bloomfield or Hughette Allen?

Surrey-Fraser Valley-Delta-Langley Green Party voters would have elected a regional MLA: maybe Bill Walsh or Kevin Purton or Bernadette Keenan?

Regional independents

The CA never had time to decide whether independent candidates should be able to run for regional seats. I‘d bet they would have said yes. Scotland uses regional MMP to elect the Scottish Parliament. Two independent candidates have won regional seats there, and two more in local seats. STV fans like the way independents can win any STV seat. But they can win any seat in Scotland too, with regional MMP.

More choices

Maybe independent Arthur Hadland would have won a regional seat in the North. Maybe independent David Marley would have won a regional seat.

The CA’s MMP model had a 3% threshold. The Conservative Party fell below that threshold in 2009, yet they got enough votes for a regional MLA in the Interior. In a real MMP election they would have been sure to exceed the 3% threshold and elect an Interior regional MLA -- their leader Wilf Hanni or Joe Cardoso or Beryl Ludwig -- and quite likely regional MLAs in other regions.

Trade-off from a province-wide model

The Green Party would have won seven seats, not six, under a perfectly proportional system with province-wide lists. Losing one seat, to get every MLA democratically accountable in a model that voters will accept, is a good trade-off.

Higher turnout

If voters knew every vote would count, more would have voted, and some would have voted differently. The Greens, for example, were so close to winning two seats on Vancouver Island and two in the Interior that in a real MMP election they would have been sure to win at least eight seats.

More women, minorities and younger candidates

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, or no young people?

The BC Green Party, for example, was very white in 2009: they had no Chinese-origin candidates in Vancouver-Richmond region, which is 33% Chinese, and only one token South Asian candidate in Surrey-Fraser Valley-Delta-Langley region, which is 17% South Asian.

Who would have been the government?

Contrary to what some Canadian newspaper headline-writers think, you cannot say the largest party will form the government. “Conservatives win!” say Canadian headline-writers even when Harper loses his bid for a majority. Compare the Times of London headline last May 7: “Britain wakes up to a hung Parliament.” No instant winner. Remember also 1985 in Ontario when Frank Miller lost his bid to win a majority. Who won? We found out only 26 days later when the Liberal-NDP Accord was signed. In most countries with more than two parties, coalitions are normal.

No “bed-sheet” ballots

Since local candidates can also be on the regional half of the ballot, voters might have had as many as ten 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.

In a 15-MLA region, suppose Party A’s voters cast 53% of the votes in the region, but elect only seven of the nine local MLAs. They also elect one regional MLA. But if that MLA dies or resigns during the legislature term, the regional candidate with the next highest votes moves into that seat. A party must run a spare. But if the seven local winners were also on the regional ballot, the party needed at least nine regional candidates, one elected, and one spare. To be safe I can see them nominating ten regional candidates.