SNTV and STV simpler and more effective than some might think

Here's a piece on SNTV and STV - both simple systems where voters do not - or do not have to - vote differently than they do under FPTP, but results are much more fair at the district and thus at the city and province and overall level.

You don't need an "advanced math degree" to conduct an STV election, as some seem to think.

And voting under Single Non-Transferable Voting is no different than voting in a FPTP election.

SNTV is an extremely simple system but produces proportional results at the district (city) level, mixed and balanced anyway even if not scientifically proportional.

Voting is the same but the difference is in how many are elected in a district.

In SNTV, there are more elected in a district (although not more in a city or a province than would be under FPTP) - and thus the results are better balanced and proportional.

STV is bit more complicated than SNTV but not as much as some think.

And in fact some STV elections produce exactly the same results as SNTV elections would have - if votes had been cast the same. (and ther are not a lot of reasons why votes would be cast differently under SNTV than STV)

The multi-member district in SNTV and STV ensures that each member elected in a city is in direct relationship with each voter in the city. Something that FPTP cannot do.

(It is possible that FPTP across a city will elect the same people as STV/SNTV elections but with only one elected in each district (a sub--part of the city) there will not be the same direct relationship between the elected member and his or her supporters (who may be in other districts and thus unable to vote for the candidate, who was elected anyway - likely with less than half the votes in the district, and sometimes as few as 17 percent of the votes cast.)

of the three systems, FPTP is a lot more accidental than STV or SNTV. Although proportional representation (looking at overall votes cast and overall parties elected) is possible under FPTP, it is likely that FPTP will not be proportional. It usually does not produce the same rep as STV or SNTV. And it is certain that about half or more of the votes will be ignored even if it does happen to produce an overall fair result.)

Even looking overall (such as at the provincial level), FPTP in many cases does not produce proportional results. (If it did, we would not need top-up seats as used in MM!)

The dis-proportionality is so bad that it happens often that a province will see one party take all the seats in the province. Say this happened in Alberta, the Conservatives taking all the AB seats. (It actually has happened), then there would be no reason to - and no un-elected Conservative AB candidates - for AB to have the 10 percent top-up seats that some seem to think the Conservative would be due..

Just an example of how top-up may be not be as easy to conduct as some make it seem.

Merely creating multi-member districts (perhaps at the city scale), with 5-10 seats in a district (although 21 seats filled at once through STV is not unknown)

and giving each voter only one vote necessarily means

- no one group can take all the seats in the city

- more than one party will be elected in the city

- any candidate that gets at least as many as votes as the fraction equal to the number of seats is ensured election.

if a city has ten seats, any candidate who receives 10 percent of the votes will be elected and there is nothing the other voters can do about it.

In fact, 9 percent (1/11th) would be assured election in a 10-seat district (if Droop quota is used) but saying a fraction equal to the number of seats is easier to explain.

And one or two will be elected at the end without even the "fraction equal to the number of seats" through being the most popular as the field of candidates thins to the number of remaining seats. (Unfortunately, the process of vote counting and the transferring of votes has long been known to be harder to explain than it is to do. A simple demonstration of STV shows its easy use. Writing it out can not.)

- voters would have direct relationship with the elected members of the city, a voter could look at any MLA or MP in the city and say that is my member and in most cases would find someone among those elected whom they support - 80 percent of so of the voters would have had their vote used to actually elect someone.

- party representation would be balanced, and the individual candidates elected would be elected due to votes placed on the individual candidate by the most voters. This is due to wide range of candidates offered to voters of a district - multiple parties and slates of each major party would contain multiple candidates.

With Single voting and multi-seat districts, the district elections would be more balanced and thus the top-up needed in a MMP-style system would involve fewer seats.

A quick study of city scale elections under SNTV

(derived from Real Lavergne's table, accessible through this: Here is a link to what I have come up with.)

The new results are based on rough calculations - simply the percentages of votes received divided by SNTV fractions.

Transfers if used would change results such as through transfers from one party to another, if vote not to be wasted (NDP to Liberal, etc.), and votes cast under more Proportional system would be different than votes cast under FPTP.

An example:

St. John's 2 seats -- no party took 66 percent of the vote in 2021 so two parties would be elected under single voting. likely Lib and NDP

In 2021 federal election, St. John's elected two Liberals

fewer Lib. seats elected under new system

2021 Lib seats new result than 2021 result

St. John's (2) Lib 2 Lib 1, NDP 1 1

Montreal (22) Lib 20 Lib 11, Cons. 3, BQ 4, NDP 4 9

Toronto (52) Lib 47 Lib 25, Cons. 16, NDP 8, Greens 1 , PPC 2 22

Ottawa (8) Lib 7 Lib 4, Cons. 4, NDP 1, Greens 0 , PPC 0 3

Winnipeg (8) Lib 4 Lib 3, Cons. 3, NDP 2, Greens 0 , PPC 0 1

Edmonton/Calgary (21) 2 Lib 4, Cons. 11, NDP 6, Greens 0 , PPC 0 (2 more)

Vancouver (13) 8 Lib 5, Cons. 3, NDP 5, Greens 0 , PPC 0 3

Looking at just these 7 places (containing a good portion of Canada's population), we see the Liberal seat count go down by 37 seats.

This would decrease the size of the top-up needed in a potential MMP system.

now with 32.6 percent of the vote the Liberals have taken 123 seats (not 160 seats as they did in 2021 federal election)

123 is 32.6 percent of 377. (377 X.326 = 123)

top-up (377-338 = 39)

so just 39 more seats would be needed (not the 153 top-up seats based on FPTP district elections that some mioght think necessary - 153 is needed to give Conservatives a lead over Liberals and/or their due share of seats based on the 34 percent of of seats they were due.)

And the voter did not do one thing differently -- he or she still went and cast one vote.

Only the vote counting was different. The votes were put together in city-wide tallies, not artificially divided into various single-member districts. Thus, different districts and the number or members elected in each district was different, while the number elected in each city and province would remain the same.

Benefits - wider range of candidates offered to voters; higher proportion of effective votes/satisfied voters; less "vote-bending" through strategic voting; better, more direct representation

I could go for a system like that.

Malta uses post-election top-up in its STV system. Because it uses thirteen 5-member districts, some dis-proportionality creeps in, despite its use of PR-STV. (This is due to having many districts, with not very high District Magnitude (low number of seats in a district).) They give a couple seats after the election to a party if it has majority of the vote but had not taken a majority of the seats. Very fair and simple.

Canada could do that too. Seldom though would a Canadian party take a majority of the votes. often under MMP, top-up is more all-encompassing than just ensuring majority votes equals majority of seats.

Canada could apply the top-up with beneficial results. (There would be some technicalities to work out as far as preserving provincial constitutional representation and who would win, though).

I suggest that more fair district elections (through multi-seat districts and Single Voting) would make top-up easier to use.

==========================

How does STV work without party lists?

STV does not derive its proportionality from being based on parties.

Instead STV derives its proportionality by its high rate of effective votes, votes actually used to elect someone - no way each party or voting block does not get about its due share of seats if 80 to 90 percent of votes are used to actually elect someone and if each elected member is elected with same or about the same number of votes, as is done under STV.

(admitted under STV the vote may be used to elect someone not the voter's first preference (but half or more of the votes are used to elect the first preference, or show first preference for a candidate who is elected ). instead under STV the vote may be used to elect someone else for whom the voter marked a preference.)

==============================

STV is really just SNTV with a bit of polish.

in STV, all or most of the eventual winners are already in winning positions in the first round, before any transfers. These would be winners under SNTV.

About transfers, most arise from elimination of least-popular candidates and the vote simply is moved to the next usable marked preference. no difficult math.

A few of the transfers in each district (never more than one less than the number of seats and often not as many as that even) are transfers of surplus votes of elected candidates. These are the complicated ones, but it can be explained that the votes are not needed to elect that person so are transferred to where someone one else may need them. This improves the proportionality and helps ensures that something like 80 percent of votes are used to elect someone.

surplus votes are the difference between the quota (usually Droop) and the number of votes held by winner.

surplus transfers are being done in  various ways in various jurisdictions:

-randomly, just by taking the appropriate number of votes  (the surplus number) from the winner's votes (from either a portion of the winner's votes or from all the votes held by the winner). again the next usable marked preference on the relevant votes are  consulted to move each vote  

- consulting the next usable preference (on just a portion of the winner's votes or from all the votes held by the winner) and deriving a micro-copy where in the surplus is composed of the same proportions 

- consulting all the next usable preferences (on just a portion of the winner's votes or from all the votes held by the winner) and composing the transfer using some of the whole votes, producing a micro-copy of the relevant votes wherein the transfer is composed of the same proportions. this a form of one of the Gregory methods.

-- consulting all the next usable preferences (on just a portion of the winner's votes or from all the votes held by the winner) and moving each vote at a transfer value, producing a micro-copy of the relevant votes wherein the transfer is composed of the same proportions. this is one of the Gregory Methods. (this actually was the way surplus transfers were to be done if STV had succeeded in BC even though no mock elections shown in educational materials  were conducted in this fashion. (un-necessarily complicated if you judge by Ireland and Malta which uses the second method listed and has done so for more than hundred years, and IMO)

so yes, surplus transfers are (or can be) complicated and are not very important really - the transfers altogether, including surplus transfers and those arising from eliminations, do not change most of the candidates who are already in winning positions in the first round. 

but they are useful  to ensure that a voter does not need to strategically vote and parties do not need to restrict the size of their slates, or at least ensures that if voter does not vote strat. it will not mean necessarily that  the vote is wasted.

===

random-ness of the first method listed is no virtue but has advantage that it is easy to do  and easy to explain. 

But equally true, the final result even with random surplus transfers is more polished than under SNTV.

so if complicated scientific transfers  are problem, go with random-ness of Cambridge (Mass) style STV (the first method listed), rather than giving up on surplus transfer altogether by going to SNTV.

(other early STVs only had transfers of surplus votes or only transfers of eliminations so those are also better than STNV, if full transfer method is problematic. )

But whether trying to explain a lesser form of STV is actually easier than explaining full STV is questionable. 

but yes voter can be like driver of a car  -- does not need to know how a oil refinery works, etc. 

basically ranked voting is like sending someone to grocery store and saying if one thing is not available, then buy something else - here's what to buy instead.

and marking only one preference (if allowed) is still likely to be enough - many votes are never transferred at all anyway.

any use of single voting in a multi-winner contest (any STV and SNTV) would be great improvement on the present FPTP.

such is -- or may be -- produced through Cumulative Voting, but CV has voter casting multiple votes and that means voter sentiment is not so calculable.

under STV or SNTV (or CV) wider range of parties would have representation and artificial regionalism and disparities would be reduced, thus reducing the polarization.

as well, attacks (personal and otherwise) would be reduced because one member would not be so significantly in the vanguard . there would be two or more leftists - and two or more rightists - in most districts so besmirching the character of one person would not deny rep. to a whole group of voters in a broad sense. under FPTP some sort of blackmail, which can be only surmised, may be employed to force members to vote certain ways. 

bribery of voters is more effective (through promises of district improvements for example) when a wafer-thin group of voters may hold the diff between win or loss and decide result.

but with quota based election system, any group whether in majority or not will have rep. if it contains a quota worth of voters. and there is nothing the other votes can do about it.

"Seat Product Model"

(from Oxford Handbook of Electoral Systems, p. 52-62)

A statistic I came across is to determine the number of  effective parties in chamber, multiply the number of total seats by average DM and find the root to the 6th power. That gives you the rough number of effective parties.

342 seats in HofC X 1 DM produces "seat product model" of 2.64, but actually five parties are rep in the HofC, although three with less than 50 seats, two with less than 20 seats.

say we went to multi-member districts of 5 DM (or range of MMDs ranging from SM 9 to 3, with average of 5)

342 X 5 = 1710     6th root number is 3.5,

so if calculation is correct -- and it works for more most democracies in the world -- we could expect about one more party in the HofC,  a slightly wider rep. than now, better reflecting the sentiment of voters even with just 5 DM districts.

and as mentioned would break up the one-party sweep of whole provinces or regions.

because if say capital city in each province were converted to some DM-5 districts or into a city-wide district, with single voting (STV or SNTV or list PR (or CV even)) , the range of members elected in that district would mean that the province as whole would have wider range of elected members.

Under Canada, constitution votes cannot cross provincial borders (IMO) so if I am correct, a MMP system in Canada would have to be regionalized like election of Scotland's Assembly, either with multi-district regions within the province or province-wide districting, but either would produce better PR than FPTP.

the Flexible district PR system (what is sometimes called urban-rural PR) includes MMDs in cities elected with STV.

The range of rep in a city MMD would help voters across the province, because a Green member elected anywhere can be a voice for Greens everywhere.

and same for a Conservative MP, Liberal MP, NDP MP, or BQ MP.

======================