(I am keeping the post below on my blog as a matter of historical reference only, and to keep the link to Carey and Hix's study.)
Stéphane Dion’s new proposal for Proportional Representation as the best voting system for Canada reflects a broader discussion: the merits of “moderate proportionality.”
Dion proposes Swedish-style pure-list PR, but with smaller regions: voters elect between three and five MPs from multi-member districts. The standard would be five-member districts. A degree of proportionality, says Dion, “moderate enough to avoid a proliferation of parties and retain the possibility of a majority government formed by a single party.”
Moderate proportional models exist
Dion is not the first to propose “moderate proportionality.” It exists in Spain, where voters elect 350 MPs from 50 multi-member districts averaging seven MPs each. It was proposed for the UK by the Jenkins Commission in 1998, which recommended a mixed proportional system with 80 regions having an average of 8.25 MPs each (some local, some regional top-up.) It was proposed for Quebec by Jean Charest’s government in 2004, when it recommended 27 regional districts, mostly with three single-member ridings and two regional MNAs. It was proposed for BC in 2004 by the BC Citizens Assembly, when it recommended a model that (after the Boundaries Commission designed the map) would have 20 districts returning a total of 85 MLAs.
The most recent academic advocates of “moderate proportionality” are Professors John Carey and Simon Hix (2009).
The “sweet spot”
They suggest a district magnitude “‘sweet spot’ in the four to eight range, where the most improvements in representativeness have already been realized but where the predicted party system fragmentation and government coalition complexity remain limited enough to allow voters to sort our responsibility for government performance and attribute credit and blame accordingly.”
They propose “low-magnitude Proportional Representation with districts in the six to eight range.” “Low-magnitude PR simultaneously fosters inclusiveness and limits the political unruliness high magnitudes invite via party system fragmentation and coalition complexity.” They find the combined probability of achieving good outcomes on all four of their variables “rises sharply moving from pure single-member districts through the low-magnitudes, peaks in the six to eight range, and then declines.” They specifically do not recommend districts as small as three. No country uses many such small districts except Ireland, and their last review concluded that the number of TDs representing a constituency should not be less than 4 unless the geographic size of such a constituency would be disproportionately large.
By contrast, the Law Commission of Canada has recommended a mixed member model “as inspired by the systems currently used in Scotland and Wales.” Scotland uses regions of 16 seats, Wales of 12. Regions of 14 seats give results twice as proportional as regions of seven seats, so many more voters can be represented by their first choice.
Districts with seven MPs at large would, under a pure-list model, mean districts containing 700,000 residents. While this might be acceptable in Canada’s nine largest metropolitan areas (which contain 51% of Canada’s population), the other 49% of us live in communities smaller than 700,000 people. Even a district magnitude of four MPs, with districts containing 400,000 residents, would only suit another 5% of us, since 44% of Canadians live in communities smaller than metropolitan Halifax’s 390,328.
Therefore, even a “moderate” model for Canada will need to be a mixed-member proportional system (known in Quebec as Mixed Compensatory), to maintain geographic representation.
A moderate MMP model
A moderate MMP model following Carey and Hix’s recommendations might have an average District Magnitude of about 7 Canada-wide; 7.1 in Quebec, 7.56 in Ontario, 6.8 in Alberta, and 7 in BC, in the new 338-MP House. Regions would mostly range from six MPs (four local, two regional compensatory), to eight MPs (five local, three regional compensatory), or ten MPs (six local, four copensatory) with exceptional three-seaters in northern BC, northwestern Ontario and Saguenay, a four-seater for northern Nova Scotia, five-seaters in New Brunswick, South Alberta, Central Alberta and Northern Alberta, and an 11-seater in Toronto Centre.
Would this exclude, for example, Green Party voters? In 2008 when the Green Party’s vote across Canada peaked, a simulation using (as Dion does) the highest remainder calculation shows this model would have done well for Green voters: they would have elected 7 of the 9 Ontario MPs they deserved, 4 in Alberta and 4 in BC.
So, although it would be less proportional than the model recommended by the Law Commission of Canada, it would be a vast improvement over Canada’s winner-take-all system, for the very reasons Dion describes.
What results would this “moderate” model produce, on the votes cast in 2011, in the new 338-MP House?
My simulation projects 145 Conservatives, 114 New Democrats, 60 Liberals, 14 Bloc, and 5 Greens. This is not quite the same as the results with perfect province-wide proportionality: 140 Conservative, 103 New Democrats, 64 Liberals, 18 Bloc, 13 Greens. That's because the smaller regions hurt the Liberals slightly. But it is a vast improvement on the projected results with Canada’s skewed winner-take-all system: 189 Conservative MPs, 108 NDP, 36 Liberals, 4 Bloc, and 1 Green. With 170 MPs needed for a majority government, an NDP-Liberal coalition would have 174 MPs.
By contrast, Dion's first model favours the top two parties.
By contrast, Dion's first model favours the top two parties.
What would Parliament look like, with a “moderately proportional” voting system?
Is “moderate proportionality” acceptable?
Fair Vote Canada prepared, in 2005, assessments of the models proposed for Quebec and BC. On Quebec, FVC said it “provides a very good foundation on which to build a fair voting system, but the current proposal must be greatly improved. With 14 regions instead of 27, a typical region could have six riding seats and four list seats. More seats per district will allow better proportionality, meaning that many more voters can be represented as they wish.” On BC, FVC endorsed BC-STV but “If BC-STV is adopted, we urge British Columbians to press for further improvements to increase proportionality and enhance diversity.” Evidently, Fair Vote Canada is not dogmatically opposed to ”moderate proportionality.” However, with four regional list seats, parties would nominate at least five regional candidates in order to have a spare for cases of death or resignation of a sitting regional MP. Nominating five candidates is, from European experience, a good minimum to get better representation of women and minorities. So regions with fewer than four regional MPs could raise equity concerns.