[EM] APR (6) Steve’s 6th dialogue with Toby

[Toby Pereira]

Steve
Before you send any reply, I just want to add something to this:
Another concern I have is that popular candidates could end up getting significantly more power than based on first preference votes. If, for example, an association makes up 2% of the electorate, they'd expect to get 2% of the power. It might be that there is one candidate in that association who is the most famous and gets more votes than anyone else. They might get 50% of the association's vote. But the other 50% might be spread amongst many different candidates, none of whom get enough to be elected. The well known candidate might still feature fairly highly in the ranks of the voters that don't put him/her top, so all the votes might still end up getting transferred to him/her. This way, someone could get 2% instead of 1% of the power. Obviously these numbers came out of the top of my head, and in practice it might not cause as much as a doubling, but then again, there might be cases where it more than doubles. This is another possible advantage of score voting, where a voter's rating of all candidates is taken into account, and it's not blind to everything below how far it gets transferred. For example, the famous candidate might get 10 out of 10 from his/her voters, but other candidates might still get 8 or 9, or possibly even 10 as well. A ranked system doesn't tell us very well how happy voters would really be if their vote was used to elect their second preference.
 
      From: Toby Pereira <tdp201b at yahoo.co.uk>
 To: steve bosworth <stevebosworth at hotmail.com>; "election-methods at lists.electorama.com" <election-methods at lists.electorama.com> 
 Sent: Sunday, 30 November 2014, 18:54
 Subject: Re: [EM] APR (6) Steve’s 6th dialogue with Toby
   
Steve (and everyone else reading)
I've cut a lot out of this to make it more concise. There's a lot we don't really disagree on, a bit of repetition and a few minor issues that probably don't warrant further discussion. As usual with my posts, my new comments have no tags at the start of each paragraph. Hopefully it's all clear.
T:  > Anyway, I'm not sure what would happen in practice with organisations being able to specify conditions for standing in their association. They could simply shut out anyone who isn't a member of the organisation to remove the competition from the ballot paper, or do you think that would turn the electorate against them?
>>>1S:  Yes, each applicant organization would want to avoid this. Each would want to elect as many MPs with the largest weighted votes as possible.  Each would try to make its application for the Primary as attractive as possible.  Each would have to consider and decide how to do this while also encouraging candidates to apply who it believes would represent it well.
If these organisation want candidates who would represent them well and they can specify their own criteria for standing in their association, then my intuition would be that they would strictly limit the candidates to their own chosen members, just like political parties decide who gets to stand for them. That's because these organisations would effectively be acting as political parties. I did suggest above that this sort of behaviour might turn the electorate against them, but I think it would likely just be standard practice for all organisations.
>>>2S:  No, the main reason is to provide an additional opportunity for each citizen to clarify their own scale of political values by being prompted to consider through which civil society “association” their own vote would be most efficiently channeled, i.e. which association is most likely to run attractive candidates.  Each association placing its list of candidates in Section A of its own Ballot is only the simplest way for each group of official electors to vote.  This is a brief way of recalling the arguments for associations and APR’s Primary repeated below in 14S:, 20S:, 21S:, & 23S:
>>>Steve’s following comparison of APR with PAL (Proportional, Accountable, Local) is based on the less than clear article at:  http://wiki.electorama.com/wiki/PAL_representation.  Consequently, he hopes that someone will be able to provide a more exact and complete description of PAL’s “assignment algorithm” and an explanation of the exact relation between “assignment” and “draft”.
I wouldn't get too worried about the exact definition of PAL, including the draft phase. I mentioned it as an alternative to APR because it's a system that someone has come up with and published on the internet. Personally I might make a few changes here and there, so I won't address all your individual disagreements with how PAL is defined. But to summarise the key points for me:
1. It is a proportional election where the whole country is treated as one region.
2. Electors would still be given "regional ballots" (where local candidates are listed), but they could still vote for candidates not on their ballot. This part would be like APR but with just geographical associations.
3. So that they do not have to rate or rank all candidates, electors can "delegate" their vote to their preferred candidate.
  >>>5)      PAL uses only “single-member” districts.
Well, the ballots would be distributed as if this were the case - there would be the same number of different ballots as candidates to be elected - but the counting process is done as a single-region proportional election. So some ballots might end up with two candidates elected, and some none.

>>>8)      Another comparative disadvantage of PAL is that each of the remaining electors’ votes are divided so that a fraction of each vote is transferred to a number of named or unnamed groups of candidates.  While these PAL arrangements may largely guarantee that each elector’s vote will mathematically not be wasted because it will help elect one or more reps, the quality of the representation resulting from PAL’s more defuse and vague relations between the agendas of each elector and their perhaps unknown fractional reps would seem less satisfactory than the closer identities that are structurally assisted by APR between each elector and his rep. The fact that each APR elector knows that his whole vote has been added to the weighted vote of his favoured rep would seem to provide a more intense personal, ideological and mutual bond between each rep and his elector.
I'm not going to discuss exactly how PAL determines which MP represents which elector because it's not something that I see as a necessary part of the process. I think a citizen should be free to choose an MP to contact as they wish on the specific area that they want to contact them about at the time. I see no need for a rigid connection between elector and MP. MPs might have their own areas of expertise and interest, and I might want to contact someone about something outside my own MP's areas of interest. In general, if I supported a party, I would probably contact my local party MP, but this wouldn't always be the case.
This is one reason why I generally speaking prefer proportional approval and score systems to STV systems. In approval/score systems, your opinion of all candidates is taken into account in the electoral process, so it should give a more accurate proportional result overall. In STV, it just starts at the top of your list and only goes down as far as it has to, so it takes less into account. For example, some people might be lucky and like a lot of MPs that their vote hasn't been transferred to. Some people might be less lucky. STV can't take this into account. Approval/score systems can do this better.
Whether an approval/score system would be appropriate for such a large election, I'm unsure at the moment, but I think it could work. For example, if I'm happy to delegate to my preferred candidate, I simply cast a vote for them. This counts as a maximum score (say 10 out of 10). This candidate has a pre-declared score for other candidates, and this is taken as my score for them. Alternatively I could score the candidates myself. Those I ignore would just score 0. My own scores could be prepared in advance in some manner to simplify the election process for the voter. For example, I could prepare a vote anonymously in advance on an official computer, and when I enter my ballot, it gives me somthing like a 10-digit code that the system saves against that ballot. That wouldn't count for anything but then I can just write in this code on my ballot paper in the election. Obviously how it would work is up for discussion and we'd have to consider all possible security concerns.
> >T: I think the main question of your method, as has been said before, is whether people would like the whole differing power thing. 
 > 
 > >>S: What people? Would not each citizen want their favourite MP to have proportionately more voting power in the Commons?
 > 
 > >T: What people? The citizens. And for your second question - not necessarily, because it also means that other people's favourite MPs might end up >with a lot of power. So in advance of any given election, I wouldn't necessarily see this as an advantage as it could work against as well as for me. Also, >as has been said, it's likely to be "celebrity" MPs that gain this extra power, so it might end up being style over substance. So I can imagine a lot of >people not being in favour of this.
 > 
 > S: It seems to me that the fundamental question is, do you want an electoral system that treats each citizen equally or not? In the above, you seem to >want one that happens to give you some advantage over others, that allows your vote mathematically to have more decision-making power than another >citizen?s vote with whom you disagree. How would you justify that?
 > 
 T: > You've missed the point here. You phrased your own question in terms of voters' wanting to have their own candidate to have lots of power - i.e. a selfish advantage for them……..
>>>5S:  Such a “want” need not be “selfish”, it might only express a desire for equal respect in an election for each elector.  In addition, some citizens “want” may be entirely altruistic, a desire for their own concept of the “common good” to be proportionately represented.
Maybe, but this part of the discussion is purely about whether people in general (who have no specific interest in electoral systems) would like the idea of differing power for MPs. And for many, the gains of having someone they like with more than the average MP's amount of power might be outweighed by the fear of someone they hate having this extra power. It's something I might consider (and see below on UKIP).
>>>6S:  Please give me your reasons for thinking that proportionate representation by any candidate within the 10% limit could be “too much”?
In the UK, we have the United Kingdom Independence Party (UKIP). UKIP is largely dominated by one man - Nigel Farage. Under your system, most of the UKIP votes would be for him, and it's possible he could even get somewhere near the 10% threshold. Some people consider him to be a nasty man and UKIP to be a nasty party. And I think all this power for one man in the party rather than spread out over several MPs could be a more "scary" state of affairs.
Of course, if we had APR as a system, the political climate in general would be different, so this might not happen in practice. So as I think Juho said, it would be interesting to see it tried out on a smaller scale. So instead of replying to all your following individual comments on the differing power part of APR, I will say that I'm open to the possibility of it, but I still have certain doubts. If I was writing my blog post again (the one that formed part of this discussion), I probably would mention APR as a system that has possible benefits.

T:> I suppose I might say that the goal is to maximise societal utility. I would generally argue for proportional systems because I think they would help to do that. I think they would help to create a more dynamic political landscape for one thing, but I think APR might hinder it in that the celebrity politicans might become hard to shift and hog the power.
>>>15S:  By what process do you have in mind for arriving at an operational definition of “societal utility”.  For myself, I see APR as providing an essential element of the wider optimal institutional arrangements for a society to make binding, evidence based and rational decisions for itself.  I would expect such tentative sovereign decisions largely both to define and to serve such “societal utility”.  The other optimal elements in my view include: a parliamentary rather than a presidential constitution, publically funded university education for all able citizens desiring it, freedom of speech, press, and association, the removal of poverty, etc.
Essentially I want laws that are fair, logical, consistent, non-arbitrary etc. When there is a debate on a piece of potential legislation, I might take a side. The side I take is unlikely to be related to where the majority of the population stand on it. That is to say that a democratic process doesn't always produce the "best" result. Of course, other systems are likely to produce even worse results. So I would want a system that is likely to produce legislation that fits my above criteria. I think a more dynamic electoral process that isn't dominated by two parties will help. So that's why I'm in favour of proportional representation. But would APR produce even better results than a system that gives all MPs equal power? I don't know.
 T:  > I'm still not 100% convinced. The other point is that while MPs can only have up to 10% of the power, a particular movement might get less than 10% representation, so it would be possible for one MP to get all the power for that movement. People might vote for just the biggest name in that movement and so they get all the votes. It doesn't necessarily mean they trust them the most. Also, it might benefit them in parliament to have allies to help them in debates.  But even if so, I don't think it would solve all the problems. A lone MP on a particular issue would welcome the support from a colleague. Or generally, a smaller number of MPs would in many cases welcome the support of a larger group even if it means their total power is the same and each individual MP's power is less.
>>>18S:  Remember that the number of MPs that each “association” (e.g. “movement” or “party”) will elect has been determined in advance of the general election by APR’s Primary.  Also, the candidates for a given association could individually or collectively “predeclare” that, if elected, they will, in effect share all their combined votes equally.
Is this correct? Assuming people don't go against their chosen association between the primary and the general election, then I agree that the proportion of power that an associaton will have has been determined in advance. An association might have 2% of the electors, but that doesn't mean they'll have 2% of the MPs, because they might all vote for one candidate in the association who then gets all the 2% of the power. So as far as I can see, they still might be left without allies in parliament.
  >>>Let me try to explain more fully why I think that APR without its Primary and its resulting associations would only be a second best APR option.  APR?s Primary and associations allow all citizens to give a proportionate extra recognition, standing and political function to the most popular organizations of civil society. >Citizens know these organizations through their work, profession, and/or daily lives and activities throughout the year. As a consequence, these >organizations (associations) have some communication and mobilization resources that are independent of the richest sections of society, celebrity, and >the mass media. These living connections help people better to know how to vote and how otherwise to participate politically in accord with their own >valued life experiences. These relations would enable many citizens to choose the sufficiently popular organizations that would become the official >electoral associations as a result of APR?s Primary. Each such citizen would rank highest the organization that accords best with his own values and >interests. 
>>>Consequently,I think APR, would also help to reduce the relative power of the rich, the media owners, and celebrity.  This is because APR’s Primary allows each citizen instead to channel their vote through the civil society organization (and its candidates) he or she sees as most consistently working to promote and protect their own abiding interests and values.  This would make mass media, celebrity, and money somewhat less important in determining how people will vote.  (Also see 21S:,23S: & 26S:)
I would certainly hope that any reasonable voting system would encourage citizens to become more engaged with the electoral process anyway. Possibly the association system would also help a bit on top of this, but it still encourages rigid thinking/lazy voting to the extent that people are encouraged to some extent to vote for candidates purely in the association that they have voted for in the primary. With a standard geographical ballot, they are obviously also presented with a set list of candidates, but they are not made to feel that they should vote for any of them on the basis that to do otherwise would be some sort of a U-turn. But this is something that I would also like to see in practice on a small scale, to see if the advantages outweigh the disadvantages (added complication of an extra election and possible encouragement of rigid thinking).
>>>24S:  If so, given all the above explanations, would you also agree that APR offers more of what you want than does PAL?
Well, PAL isn't the only option for me (as discussed above). There are some good arguments for differing MP power and for APR's primaries that you've made, but as you can see I still have some concerns. I'd like to see it in practice. But if I had to pick an election system right now, I probably would still on balance go against differing power and associations!
Toby

   
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20141201/6d60fe3e/attachment-0001.htm>