On the topic of semi-proportional outcomes, we should note that all PR systems reap semi-proportional results. Technically speaking, the only exactly proportional system would be one where every voter is elected.
Even party-list PR does. Where there is a threshold of 5 percent, as much as 20 percent of the votes are disregarded just out of the gate. Such as Estonia in its last election. Some waste occurs in New Zealand MMP elections as well.
There will always be rounding off. When you have 338 representatives representing 18M voters, such as we have in Canada's federal elections, there is considerable rounding off. Every seat represents 53,000 votes on the average, so anything less than about 26,000 is less than half so round-down-able and lose-able. That is taking the country as a whole. The use of cities or sub-provincial regions as districts or even provinces as districts would make a PR system less proportional still, although it would be much, much more proportional than the present system.
I generally say STV produces mixed roughly-proportional representation at the district level.
Representation in a STV district is proportional relatively-speaking -
- the most popular party gets more seats than the next one, or at least no fewer, and
- no party is popular enough to win all the seats in a district, and
- a small party gets at least one seat if one of its candidates receives quota, or if the supporters of its candidates vote along party lines and together add up to quota. (or at least if it suffers no more "leakage" than another party, or if it does not have quota itself but attracts votes through transfers from eliminated candidates of other parties).
This proportionality is produced not through examination of party vote tallies (they are used just as test or measure of proportionality after the fact) but through the magic of single vote in multi-member district and transferable votes.
A flexible system that
- allows Independents or party candidates to be elected,
- elects the most popular candidate of each party that is awarded seats, and
- even guarantees representation of a particular sub-district. If the voters there add up to at least quota and concentrate their choices on their locality's candidates, they will elect at least one representative. There is nothing the voters elsewhere could do about it. This should satisfy those worried about the loss of local representation under multi-member districts.
STV should really be called proportionate representation according to one of its Canadian proponents in the late 1800s. This was Sir Richard Cartwright, the longest serving MP in Canadian history - 43 years in the House. True fact. (Even legendary Mackenzie King served only 32 years.)
SNTV probably gets us about the same results as STV, although without its polish and with more waste of votes.
Like its proportionality - somewhere between FPTP and STV - the amount of waste of votes under SNTV is also in that same middle. (I think)
The amount of strategic voting under SNTV would also be in the middle zone compared to FPTP (massive amount) and STV (almost unknown).
People wanting not to waste their vote under SNTV would place it with a candidate expected to be among the leaders to the same number as the number of seats, so wider choice than under FPTP, where votes worrying about potentially wasting their vote often feel they have got to vote for only one of two candidates. Under SNTV voters would have choice of about 3 to 7, more or less, choices, whatever the seat count is. So a funnel with a wider mouth.
Oddly the proportionality of SNTV is thought to vary from perfect in both directions.
Some say that large parties suffer due to possibility of a massively popular candidate taking more votes and thus "starving" the other candidates of the party. (I have read of instance under STV where town leaders advised voters to vote for a lesser-candidate, hoping to ensure his election as well as the most certain winner. (This was unnecessary under STV but they apparently did it anyway.) The lesser won great number of votes and easily took a seat; the "surefire winner" barely squeeked in. This is taken as case study in how you should not try to play the system but just vote for whom you want elected.)
But the ACE Electoral Knowledge Network website says that large parties benefit from SNTV. This should be read with caution. They are apparently comparing SNTV to an utopian perfectly proportional system. It howls with derision that in Jordan a party with 48 percent of the vote would get 55 percent of the seats under SNTV. Heck in Canada's FPTP elections, a party with 48 percent of the vote (or even less) gets 90 percent of the seats (or even more) under FPTP.
Their point is that small parties whose supporters are spread out would suffer - and therefore that large parties benefit. But even that is comparing it to at-large PR elections. The large districts under SNTV would give these spread-out parties a better chance than they would have under FPTP with its multitude of small districts breaking up and dividing the electorate.
I do think SNTV would work better than our FPTP, and STV even better.
Thanks for reading.
=============================
Comments