Tuesday, April 12, 2011

Please explain: Preferential and proportional voting

One of the criticisms made against proportional voting is that it makes it easier for extremists to gain a seat in parliament since a successful candidate only needs to secure a relatively small percentage of the vote. Indeed, it had been looking like Pauline Hanson, the extremist Australians most love to hate, was going to get a seat in the NSW Upper House after the recent election.

However, it was not to be, because proportional voting in Australian Upper House elections is combined with preferential voting, and so even though Hanson won more primary votes than the two other candidates with whom she was competing for the final two seats, on preferences, they both overtook her.

Preferential voting prevents extremist candidates from winning in races where multiple candidates split the vote, since it allows voters the chance to indicate who is their last preference, as well as their first. UK voters, vote "yes" to electoral reform on 11th May.

7 comments:

byron smith said...

Speaking of electoral reform, here is a suggestion I hadn't heard before.

Anonymous said...

Personally, I think our preferential system, at least the way it is at the moment, is grossly unfair. Take the last NSW state election as an example. The seat of Balmain has been won by the Green candidate, despite him coming second with 14,019 primary votes to the Liberal candidate's 14,860. Why do we have a system that works out why the person who comes second should be made the winner?
The other problem with our preferential system is that not everyone actually gets preferences. The 14,860 voters (ie: the majority) who voted Liberal are not given a second choice, despite the fact that they won't be getting their first choice no less than those who voted for the lowest 6 candidates.
No system is going to give everyone what they want, but Balmain is now represented by a member who was not the choice of over 70% of voters. And as we currently see in the Marrickville council, Greens enjoy pursuing their own agendas, regardless of how idiotic and ill-supported they are. -JC

byron smith said...

JC - The preferential system is more fair, rather than less fair. Rather than asking purely about first preferences, it asks about multiple preferences and so ensures that the winner has the goodwill of at least half the electorate. Most people I know don't have a single opinion, but a variety of opinions about different parties and candidates and preferential voting lets you not only say who you want, but also who you don't want.

To take your example: in Balmain, there were more people who didn't want the Liberal candidate than who didn't want the Greens candidate. The 14,860 voters (who, by the way, were not a majority, or there wouldn't have been more than a single round) who voted for the Liberal candidate did indeed get to express a second preference, which was for the Liberal candidate! That is, preferential voting is basically the same as a run-off election (used in many countries), but with all the rounds occurring at once.

Balmain is now represented by a member who was not the choice of over 70% of voters.
Not correct, unless you restrict "choice" to first preference. A majority chose the victor over any given alternative. This cannot be said for first past the post. If Balmain had been conducted with FPTP, then your statement would be true (Balmain being represented by a member who was not the choice of over 70% of voters). Such a situation is precisely what preferential voting is designed to achieve.

If 100 people are voting for A, B and C, and 35 want A (but definitely don't want B), 33 want B (but definitely don't want A) and 32 want C (but also definitely don't want A), then making A to be the winner would be strange, since 65% would prefer B to A.

FPTP only makes sense when there are two options.

byron smith said...

(BTW, I don't think the preferential system is perfect - far from it - just better than FPTP for the context of an election with more than two parties).

Anonymous said...

(I’m not meaning to be argumentative, but) I still think our current preferential system is typical of the attempt to please everyone while satisfying no one.
To say that the second preference of those voting for the top two candidates is the same as their first is not really letting them answer the question “If you can’t have the one you want – who would you choose second?” Having worked on several elections, many of the preferences distributed to the top two candidates are actually the voter’s 7th or 8th choice (or lower) which is not so much ensuring the winner has the goodwill of the majority as it is asking: “Who do you hate the least?” The whole concept of ‘majority rules’ breaks down when you have more than two choices; 50% is no longer a magical figure.
It could well be that the compromise between “Who I want” and “Who I hate” means that the third placed candidate is the most palatable. In the case of your A, B & C example, it would also be strange to make B the winner since 67% prefer C to B.
Theory and practice can be vastly different things. Since preferences are optional in NSW, it is possible for no candidate to get an outright majority. Getting back to A, B & C – how would it affect the goodwill factor if the entire vote for B gave their second preference to A? The preferential system would not consider this.
Ultimately, I just wish I had a list of politicians I support, rather than a list of those I can’t stand. And I would like a system that separates who represents my area from who legislates for everyone. -JC

byron smith said...

Thanks for pushing back. I admit that there are some very specific situations in which preferential gives a perverse outcome. My claim is simply that these are relatively rare in comparison with FPTP, which gives them very frequently in an electorate with a three or more cornered race.

Yes, preferential may end up being (for some people) "which party/candidate do I hate least?", but this is still a valid piece of information.

Optional preferential (as in NSW, and as in the proposed UK AV) is superior in my opinion to compulsory preferential, even though it can result in no candidate receiving 50% of the vote. I am not aware of any actual cases in which that has occurred during a NSW election (do you know of any?). As you say, theory and practice can be different things. Some of the problems with preferential voting are usually only hypothetical. There is nothing hypothetical about the fact that the most recent UK election saw two-thirds of electorates won by candidates with less than 50% of the vote, many with less than 40%. Indeed, when combined with falling voter turnout, not a single constituency in 2005 returned an MP with 50% of eligible voters in support.

Antony Green points out an electorate in the recent NSW election where FPTP would have given a seat to Labor with 26.3% of the vote, when there was clearly a huge majority voting for the Coalition. He also points out that no NSW seat has ever been awarded to a candidate coming third on first preferences.

AV is not perfect (no system is), but in the circumstances faced by the UK today, it is superior to FPTP.

Anonymous said...

I didn’t know the UK is about to hold a referendum on preferential voting, so I was unaware how topical this is until your post of May 4. Maybe my first claim that preferential voting is “grossly” unfair was a slight exaggeration.
What I like about the preferential system is that it does allow voters to answer three questions:
1. Who do you want to win?
2. If you can’t have your first choice, whom else would you like?
3. Who do you not want to win? (The voter’s last placed candidate is the only one who can’t end up with their vote, or their blank candidates under optional preferences.)
The unfair part is that the method of distributing preferences means not all voters actually get their answers to all three questions considered, or equally regarded. In fact, it is the largest two groups of first preference voters who do not get to answer questions 2 and 3, and can have another voter’s 7th preference held against their 1st. All this means the resulting ‘goodwill’ that flows to the eventual winner is a skewed version of what was actually expressed by the entire electorate.
No voting system itself has any impact on the quality of the candidates elected. Any system can result in the shutting out of a ‘good’ extremist or the shunting in of a ‘bad’ one. -JC