(part 4 of a multi-blog research essay)
(This is the final part)
What is the ultimate size of the multi-seat district?
The entire state is the largest possible constituency for the election of state legislators in the U.S. Just as a province as a whole is the largest size possible for election of provincial legislators in Canada.
PEI is about the only province where this is possible, with 27 MLAs. Other provinces have many more MLAs, from Ontario's 124 MLAs to Newfoundland and Labrador's 40.
As the number of seats increases, so does proportionality, but so also does the size of the physical ballot and the burden of counting votes. So the number of seats is set through a balance of the two opposing indices.
MM districts produce four benefits:
longer legislative tenure (SMD produced this as well)
less gerrymandering (SMD produced this as well)
more representative legislature*
multi-party system*
* Silva points out that the extent of these benefits depends on the number of members apportioned to each district and on the kind of electoral system used in the district. (p. 509)
Seats in the Provinces
NFLD 40 MLAs 7 MPs
PEI 27 MLAs 4 MPs
Nova Scotia 55 MLAs 11 MPs
NB 49 MLAs 10 MPs
Quebec 67 MLAs 78 MPs
Ontario 124 MLAs 121 MPs
Manitoba 57 MLAs 14 MPs
Sask 61 MLAs 14 MPs
Alberta 87 MLAs 34 MPs
BC 87 MLAs 42 MPs
total 654 MLAs 338 MPs
Can province-wide districts work?
As a precedent of how large a multi-seat district can be, we can look at New South Wales (Australia).
In its 2019 election of members of the legislative council, NSW used STV in one multi-member district to elect 21 members. 4.5M votes were cast in NSW in this election.
That number of votes cat is more than are cast in any province in Alberta.
The number of seats filled through STV in that one at-large district in NSW is more than are cast federally in six provinces. so there would be no hold-up to using STV in a province-wide district at least in those six provinces, if we accept that the 2019 at-large STV election was a success.
But the NSW election does not give us precedent to using STV in a province-wide election for provincial elections as there are no provinces where 21 or fewer seats are filled in provincial elections.
But elections in even the largest province, Ontario, could be theoretically held using just six districts. That is, if we were to part from the current belief in single-member representation.
ONTARIO
About half of Ontario's population is within the Greater Toronto Area, currently divided into 52 federal ridings, while the rest of the population is enclosed in the other 69 ridings. [More on this in my blog "Rural seats are being lost - We need more members overall"]
And the provincial district structure is likely much the same.
So federally
GTA could be grouped into three districts of 14 seats each (quota would be about 7 percent)
the rest of the province:
6 of the largest ridings would remain single-member districts (the large size of a MM district making cross-district travel very arduous)*
all other ridings (63 of them) could be grouped into three districts. (Quota would be about 4 percent)
or say six districts of about 10 seats each (Quota would be about 9 percent).
The large ballots that would result from high District Magnitude (large number of open seats) would not be too daunting if candidates' names are arranged by party, and the burden imposed on voters voting under preferential voting would not be too heavy if voters had liberty to mark only as many back-up preferences as they desire.
If the same districts were used for both federal and provincial elections, or if one level moves to M districts before the other, if the MM districts were composed of a number of the units of the other, a natural relationship between members and voters and a local culture and network could emerge. Of all the Canadian provinces, this is only possible in Ontario where the number of provincial seats and federal seats are roughly the same. In all other provinces there are many more provincial members than federal members.
Under STV, despite large MM districts, the possibility for local representation could be assured by the knowledge that any locality or corner of the multi-seat riding would elect a local representative if the voters there marked their preferences consistently for local candidates and if they have quota. And elected members would be elected by votes within each district, with responsibility basically only to represent the residents of that district.
* Districts that remain as single-member districts could switch to Alternative Voting. Voters there would cast preferential votes same as voters in the other regions of the province, in order that voters everywhere use the same voting system.That way all voters across the province would cast preferential ballots - marking first preference and back-up preferences if voter desires. Having the same voting system in use across the province would simplify the broadcast of instructions to voters.
========================================================
Multi-member districts in our future?
Taking Manitoba as a project, how could fair representation be achieved through multiple-member districts?
WINNIPEG -- Under FPTP, in the last Manitoba provincial election (2019), there was considerable dis-proportional mis-representation.
A few multi-seat districts would easily cover the city and with a fair voting system, would produce representation much fairer than that produced in 2019.
Winnipeg
Overall Winnipeg (32 seats): 15 Cons 14 NDP 3 Lib
overall party vote percentages: Cons 37 percent NDP 36 percent Lib 19 percent
number of winners that won with majority of the district vote: 14
number of winners that won with minority of the district vote: 18
Winner's percentage of the vote (high and low): 62 p.c. (Point Douglas) and 38 p.c. (McPhillips)
Number of winner's votes (high and low ): 6075 (Robin) and 2536 (Burrows)
Effective votes by party: Cons 69,000 (67 percent of Cons votes)
NDP 51,000 (51 percent of NDP votes)
Liberal 13,000 (26 percent of Lib votes)
Overall effective votes (votes cast for candidates who were elected):
133,000 votes 48 percent of valid votes
Overall wasted votes (votes cast for candidates who were not elected):
142,000 votes 52 percent of valid votes
But PR would be produced if Winnipeg was divided into say two multi-seat districts:
one of 15 seats Quota would be about 7 percent
one of 17 seats Quota would be about 6 percent.
An odd number of seats in the district is preferred as it allows a party with a slight majority to have a majority of seats easily.
Under STV, about 80 percent or more of the voters in a district would see their vote actually used to elect someone. This is a much higher figure than FPTP. In Winnipeg less than half the voters saw their vote used to elect someone.
Although Winnipeg is a large city (by Canada standards), elected members could easily cross the city if it was one district and easily cover half the city.
Each district, even those covering about half the city, would still be smaller in geographic area than even the smallest rural district. And there would be multiple members to do the job, not just one as under FPTP.
The whole metropolitan area of Winnipeg is only 464 sq. kms. in size.
The average for provincial districts in Manitoba is 11,000 sq. kms.
Leaving out Winnipeg, pretty much all of Manitoba is divided into 25 districts, each represented by just one MLA. so each district outside of Winnipeg on average covers almost 26,000 sq. kms.
(Back in 1899 in the age of horse and buggy, 37 districts covered Manitoba outside of Winnipeg.
If there are large rural districts, which there are, it is partly due to the number of MLAs outside Winnipeg decreasing in number as Winnipeg took a larger portion of the MLAs elected each time.
The same holds true of Alberta. Fewer MLAs represent the area outside of the major cities today as compared to 1917, back when automobiles were made of tin and most roads were dirt.
1917 52 MLAs MLAs outside of Calgary and Edmonton
2019 41 MLAs outside of Calgary and Edmonton
We have more people and more money compared to 1899 or 1917, but only slightly larger number of MLAs overall and fewer MLAs for rural areas.
A basic fairness would be created by researching the largest number of seats for the rural portion of the province ever in the province's history and determining the number of voters in an average district (by diving the total rural pop today by that number of MLAs) and using this as a guide as to how many seats Winnipeg should get, with say an average district in Winnipeg having a third more voters than a rural district.
With Winnipeg using multi-seat districts, there would be flexibility in the number of members the city would have.
The districts could based on the city's corporate limits and on natural lines within the city, such as rivers.
With rural districts being stable, there too organic districts could be drawn. They could incorporate natural boundaries and geographic features.
The districts' boundaries would stay the same. Just the number of seats could change in subsequent re-distributions.
The total number of MLAs would change as population grows. But the shape and size of the districts would stay the same.
Grouping districts should be easily do-able within Winnipeg itself. Certainly the geographic size of new MM districts within Winnipeg should not be a problem.
At the time when rural districts are stabilized, it might be easy to also create multi-seat districts in rural regions. Grouping districts outside Winnipeg into MM districts may be difficult with the existing number of MLAs in the province (57).
Perhaps a small increase in the number of MLAs would ease this process. An increase in the overall number of elected members is sometimes par for the course when moving to a more fair electoral system. And such an increase could help when creating MM rural districts.
When New Zealand moved to MMP, the overall number of elected members was increased by 20 percent.
In Canada too such an increase may be considered a worthwhile trade-off for more fair and balanced representation.
That size of increase would be about 60 MPs in federal elections.
In Manitoba it would be an increase of 10 MLAs.
To balance urban and rural representation, 5 could be added to Winnipeg, where elections would be held in MM districts covering half or smaller portions of the city.
Five additional members could be added outside Manitoba.
We could take the ten rural districts with the largest populations. (Likely these would be the ones smallest in geographic size).
Adjoining neighbours could be paired and each pair (now a MM district) could be given an extra seat.
In the new 3-seat districts that would result, under STV any candidate who received 25 percent of the vote would be elected.
Only if a party had 75 percent of the votes would it take all the seats in a district.
If MM and single voting was brought in within Winnipeg and in ten populous districts outside Winnipeg, the election of representatives for more than half the population of the province would thus be held under PR. Under such a scheme, the voters that are currently represented by 42 of the province's 57 MLAs would elect their members through PR.
Under the scheme, those voters would be represented by 52 of the province's 67 MLAs.
And the depth and nature of that representation would be of a much higher quality than currently under FPTP.
Under such a PR scheme, Liberal and NDP voters outside Winnipeg who are now generally under-represented would see their views better represented in the Assembly than under the present FPTP.
Currently in Manitoba's rural districts, Conservative candidates are winning seats with barely more than half of the vote while as much as 45 percent of the voters get no representation at all. With 3-seat districts and STV used, two or more parties would be elected in many districts.
Brandon East and Brandon West are natural choices for paired districts.
In the 2019 election, Brandon was divided into those two districts and FPTP was used to fill the two seats. 7600 Conservative voters in those two districts took both seats, while 4000 NDP, 1400 Liberal and 700 Green voters got no representation at all.
Under STV, you would expect 80 percent of the votes to be used to elect someone. With Winnipeg's 270,000 votes and about 80,000 votes in the ten current districts outside Winnipeg that would be made into MM districts, we would have 350,000 voters out of the province's total of 479,000 (2019) electing under PR and 280,000 would see their vote used to elect someone.
In the 2019 FPTP election, in those same parts of the province only about 180,000 saw their vote used to elect someone.
But multi-member districts by themselves are no solution to dis-proportional misrepresentation.
The flexibility produced by having MM districts means that if evil is intended, officials can draw a border that best gerrymanders against the opposing party then apportion whatever number of seats the size of the population within warrants.
Small parties can take a seat or two under single-member districts when they might be totally shut out if large MM districts are used - if an unfair voting system is used.
But MM districts means that some balance is easily achieved - by merely restricting each voter to only casting one vote. X voting, the type of voting voters are used to, could be used in MM districts and would often produce mixed representation. It prevents one-party sweeps unless that party has a great proportion of the vote or opposing parties suffer from vote splitting (perhaps due to running more candidates than their vote portions warrant).
Even if that vote is not transferable, the casting of a single non-transferable vote in a MM district would often produce mixed, balanced representation.
STV uses transferable votes and is more dependable than SNTV.
The use of non-transferable votes in SNTV means that parties more popular within a district may get fewer (or no) seats than a less-popular party. The more popular party usually does get more seats but it is not to be depended on. But the use of SNTV in Vanuatu shows that MM districts do produce mixed representation. Mixed representation is more fair than one-party sweeps, the usual outcome of FPTP elections across multi-district regions - or even provinces - or even multi-province regions.
(Information on Vanuatu's 2020 SNTV election can be found at
The use of transferable votes (STV) will ensure more dependable and proportional results. The representation elected in each MM district will be dependably more proportional to the votes cast in the district than the result under SNTV. Parties with less votes than another party will not get more seats than that other party. At least that was the record in Edmonton and Calgary in the 30-plus years that STV was used (based on first count tallies).
Are transferable votes necessary?
The SNTV and STV systems in some cases elect exactly the same people - and seldom do STV final results vary much from the results that would be produced by SNTV.
The drawbacks of preferential voting that is used under STV is only a bit more burden put on voters - although voting along a party line is easy enough, but perhaps more significantly there is the psychological challenge of switching the voting system, where any change appears to be a hard-fought battle.
In conclusion,
Multi-seat districts are key to district-level PR voting systems.
Having more than one member in a district may seem revolutionary and risky - but I hope that the evidence presented in this essay helps the reader see that MM districts are not so out-there.
Tom Monto, Edmonton Jan 2022
=========================================================
Later remarks:
Some seem to have idea that we inherited the FPTP system from Britain, and that was the way it was - the only way it was and now those who fight for Proportional representation are trying to bring in something that is totally new and totally un-tried.
But starting at the time of Confederation (or even earlier), multi-member districts, not the single-member districts commonly associated with FPTP, were in wide use.
The two original Atlantic provinces at 1867 - NB and NS - and PEI, when it joined in 1873, all elected most or all of their MLAs in multi-member districts, not in single-member districts.
BC MLAs were almost all elected in MM districts in its first provincial election (1871).
and Multi-member district(s) were used in every federal election prior to 1968, starting with the Confederation election of 1867. The last federal multi-member district was used in Canada in 1968.
From 1867 to 1968, Canadian federal elections used one or more two-seat districts.
The two-member ridings were:
Ottawa 1867 to 1933
Halifax 1867 to 1968 (the last federal riding to have multiple members)
Victoria, BC 1871-1904 (1871 under the name Victoria District)
Cape Breton, N.S. 1872-1903
Pictou, N.S.1872-1903
St. John City and County, N.B. 1872-1896
Hamilton, Ontario 1872-1903
West Toronto (Toronto West) 1892-1904
In PEI: King's County 1873-1896;
Queen's County 1873-1896;
Prince County 1873-1896.
So as you see, multi-members districts were quite common in 19th Century Canadian elections. And through most of the 1900s as well.
(MMDs in old Britain)
And British elections too used MM districts back in those days.
For instance, they used Limited Voting to some degree, co-incidentally starting at the time of Canada's Confederation. The Limited Voting system "was applied in England to constituencies returning more than two members from 1867 to 1885." [Footnote states "the LV was applied to 13 constituencies each returning three or four members and included Glasgow, Birmingham and the City of London." (Phillips, Challenges to the Voting System, 1867-1974, (p. 17) available online: https://ir.lib.uwo.ca/cgi/viewcontent.cgi?article=1860&context=digitizedtheses)
British elections might also have used Block Voting in other districts or times as well.
The dominance of FPTP, the idea that Canada and Britain always used it before and now pro-PR reformers are trying to bring in something completely new and never tried - like the myth that no suitable alternative exists - is a myth that has wide currency but is not based on truth. It seems those who do not know their history are doomed to repeat... doomed to repeatedly elect un-representative governments!
==============
Effect of MMD may be no improvement over SMDs
Large district magnitudes increase the chance for diverse walks of life and minority groups to be elected. However, it is not synonymous with proportional representation. The use of "general ticket voting" prevents the multiple-member representation of the district from being mixed and balanced.
Where list PR is used in the district, a closed list PR method gives the party machine, not the voters, the power to arrange the candidates on the party list. In this case, a large district magnitude helps minorities only if the party machine of any party chooses to include them. In a multi-member district where general ticket voting is not used, there is a natural impetus for a party to open itself to minority voters, if they have enough numbers to be significant, due to the competitive environment produced by the electoral system.
=============================================
(THE END)
If you missed reading the first three parts of this multi-blog essay, here's the links:
=================================
Comments