Surpluses are treated differently from eliminated candidates in the election

JohnCondon said:
... If you think eliminated votes and surplus votes should be transferred whenever they're transferable, then there is no inconsistency.
Click to expand...
But the point is that in the case of a surplus, transferable votes are not transferable in whole but in some proportion. The Scottish/Australians choose the proportion to the original votes, the Irish system chooses the proportion to be to those original votes that expressed a preference. As I point out this effectively treats those who did not express a preference the same proportion of preferences as those that did do so.
It's a complicated and well thought out system and presumably they had the benefit of knowing the "correct" Australian approach. So one assumes the legislators deliberately chose this "illogical" version. It favours the larger parties so that would not be a huge surprise.
Having said that, it is hardly a big scandal and the occasions when someone who would have got elected on the "correct" system but failed to do so on the Irish system must be rare.
 
Last edited:

Non-transferable votes are never transferred, in surplus or in eliminations. I get how it can feel like they are, but in reality they are not.

What happens is the returning officer must arrange the votes into sub-parcels, setting the non transferables to one side. If a candidate has surplus over the quota, then transfers are made from the piles of transferables, either one-for-one, or proportionally, this is the legislation. Only transferables are ever transferred. The non-transferables do contribute towards the candidate's total vote though, and this helps to produce the surplus that is then distributed based on the transferables. Non-transferables are not transferred. That is a simple fact. If there are 1000 non-transferable, 1 transferable and 1 surplus, then just 1 is transferred and it is the transferable vote.

To transfer a vote is to move it from the elected/eliminated pile into the continuing candidates' pile, in real terms and according to law. In all the elections and counts that happened last week, a non-transferable vote was never transferred. E.g. Baker's 45 non-transferable were retained in her quantum of votes to match the quota, and only transferable votes were eligible to be transferred.

Neither Scotland nor Ireland are "right" or "wrong" on this, they just have different perspectives. To get to the heart of this question, we need to ask what truly is a surplus and what is the point of it, what is the point of transferring votes, what is the point of filling out a ballot the whole way down. What is the point of a proportional representation system. If we try to properly deal with the issue in this thread we could fill a book with arguments for each side. I'm not going to delve into that but the issue is more complex than simple right or wrong.

P.S. Two things, to me, are very disappointing about the Irish system. In surplus, on counts 2+, only the votes in the last parcel that took the candidate over the surplus are counted in further passing on the surplus. And we have effectively have a random lottery of selecting the votes that are physically transferred. That randomness, to me, is unacceptable.
 
Brendan Burgess said:
I don't see the false equivalence and obviously the people who designed the far more logical Scottish and Australian systems don't see it either.
Click to expand...

More logical but less representative?

In post 37 (real life example), 1306 surplus to be distributed.
The “logical” approach is to select 1286 to transfer and 20 to put in the bin, to ensure the proportions of 2nd/no further pref, are maintained. The mathematical proportions.

However, the representative view would be not to select 20 dead ballots, from voters who no longer care about the outcome but instead to ascertain the views of 20 voters who do have a preference. So, logically, no votes are wasted unlike the “logical” approach. This is entirely consistent with the principles of the STV system, your vote stays with only one candidate unless it is not needed by that candidate (due to already having sufficient votes I.e. election or elimination).

Looking at our example again, this does result in an uplift of a handful of votes for everyone. Dockrell receives 0.45% more votes relative to their share of the total vote. But the Total Vote as the denominator is what is in question. The total vote consists of people who are engaged in the process (transferable votes I.e. useful for determining future outcomes) and people who have chosen to disengage from the process (non-transferable I.e. useless for determining the future outcome). In determining the future view do you ascertain the average view or do you ascertain the view of only the people who are actually interested in the future outcome?

The rational and logic of the Irish system is completely valid. Yes it is different from the the Scottish and Australian system. The Irish system is a product of the early 20th century, the Australian system of the 40’s and the Scottish system of the 21st century. There may be other reasons why their nuances were adopted, namely to reduce complexity, aid machine counting etc. But they have sacrificed representativeness for reduced complexity. (Marginally!) The Irish system is neither right nor wrong.
 
JohnCondon said:
Non-transferable votes are never transferred, in surplus or in eliminations.
Click to expand...
Nobody is arguing that non transferable votes are physically transferred.
Another example.
Quota 8,000 A gets 10,000 votes. 4,000 are transferable. Under the Irish system 50% i.e. 2,000 of these would be transferred. The Scottish/Australians would say 80% of those who voted for A did want to transfer their surplus and so they would argue (correctly IMHO) opinion that only 400 of the transferable votes should be transferred.
Agreed. But our foray into electronic voting did not have a happy ending.
 
Yes and to be consistent the "dead ballots" of eliminations should be "allocated" to those who still care about the election, most elected candidates would then get a full quota.
Now maybe the designers of our system (which you inform us was something of the original, thanks for that) would have wanted it that way but since the physical process involves parcels of whole yet single papers it could not work in practice.

I very much doubt that the later versions (Australian/Scottish) went for the "logical" approach because it was administratively easier or more computer friendly. One presumes they saw it as a correction to the earlier version. But I agree this does not amount to a constitutional crisis.
 
Last edited:
Hi John

Let me try with an exaggerated example to show what is wrong with the Irish system.

A gets 1,000 votes
The quota is 600
So her surplus is 400

Scenario 1 - all votes have a 2nd preference
Transfer Value = 0.4 ( 400/1,000)

A's second preferences are as follows

B: 500
C: 500

Both get 200 each

Scenario 2 - only 50% express a second preference
A's second preferences are as follows:
B: 500
C none
D non-transferable : 500

Now the TV = surplus/transferable votes = 400/500 = .8

A's votes are transferred
B 500*.8 = 400


This strikes me as clearly unfair. B gets half the 2nd preferences but all the surplus.

Scenario 2 - Scotland

The Transfer Value would be Surplus/Total Votes = 400/1000 = .4
So B would get 500 * .4 = 200
 
JohnCondon said:
The non-transferables do contribute towards the candidate's total vote though, and this helps to produce the surplus
Click to expand...

This is the point we are making. The non-transferables contribute towards another candidate's total vote and they should not do so.

"the non-transferable are transferred" strikes me as a good summary of this.
 
Itchy said:
However, the representative view would be not to select 20 dead ballots, from voters who no longer care about the outcome but instead to ascertain the views of 20 voters who do have a preference. So, logically, no votes are wasted unlike the
Click to expand...

I don't get this at all. If I choose not to vote, you should not be allowed to vote for me, so that my vote is not wasted.

Because the numbers are small, you don't appreciate the unfairness of the system.

Does the most recent example I gave of someone getting 50% of A's transfers but all of their surplus not illustrate that this plain wrong? Not just different to Scotland. But wrong.

Brendan
 
How would you naturally go about distributing a surplus with a computer program?
So take our example. Quota 8,000. Votes for A 10,000. Surplus 2,000.
You would say that only 80% of the 10,000 was needed and so there is still 20% of a vote that should live on. The program would then mark down all 10,000 votes to a value of .2.
Or even without a computer the natural approach would be to somehow select (at random) 2,000 of those votes and then transfer them or not as the case may be.
Can't see any logic for our system at all.
And Brendan has shown clearly that the well known economic commentator Cormac Lucey has been a "victim" of this illogic.
 
Last edited:
Duke of Marmalade said:
Yes and to be consistent the "dead ballots" of eliminations should be "allocated" to those who still care about the election, most elected candidates would then get a full quota.
Click to expand...
Brendan Burgess said:
Does the most recent example I gave of someone getting 50% of A's transfers but all of their surplus not illustrate that this plain wrong?
Click to expand...

Again, That’s an absurdity to do that. If the votes for Loo La No. 1 at the bottom of pack, say 400 votes are distributed as a result of an elimination with 50% non-effective and 50% for Loo La No. 2, it would be ridiculous to give 400 votes to LL2 when there are physically only 200 ballots with the number 2 written on them. If I was a voter for someone else I would obviously have a grievance.

Per Brendan’s example, If they are distributed as part of a surplus, the difference is, I physically have 400 ballots with the Number 2 written on them.

Therefore it’s a false equivalence. It’s not inconsistent to apply different process when different underlying scenarios occur.

Brendan Burgess said:
I don't get this at all. If I choose not to vote, you should not be allowed to vote for me, so that my vote is not wasted.

Because the numbers are small, you don't appreciate the unfairness of the system.
Click to expand...
Which voter is being unfairly treated? If it’s the indifferent voter that I’ve already mentioned, could you address that point?
I fully accept the TV maths by the way. I’m more interested in the principles. Which principle of the STV/PR system is being violated?

Can I attempt a different analogy? Say I’m doing a survey in my estate of 20 houses, with two parts. I want to find out if we’d like a crèche in the estate (part 1) and if so what facilities it should have (part 2).

So part 1, I ask all 20 and great were all on board. Now a particular nuance of this group is have half have kids and half don’t and therefore don’t really care about what facilities are on offer.

So part 2, I want to find out what’s important in terms of facilities. Now the half that don’t care, did not complete part 2. They like the idea of having a crèche but have no interest in what colour the walls are.

So say one response want late pick ups, Two responses want early starts, six responses want after school places and one wants something else/not defined.

Which is a more useful summary of the survey results?
A) 100% want a crèche, 5% want late pick ups, 10% want early starts, 30% want after school places, 55% something else/not defined.
B) 100% want a crèche, of those who care, 10% want late pick ups, 20% want early starts and 60% want after school places, 10% something else/not defined.

In the former, we’ve examined the total view for each part, in the latter we’ve only examined the view of those who have indicated their interest in the outcome of each part.

Surely you can see that Summary B is more useful & representative of the views of the people rather than the mathematical presentation?

The subtlety of the Irish system is that in the determining the desired future outcome, it gives higher weight to votes that are interested in that future outcome, rather than the Scottish system which equal weights voters who do and voters who do not have an interest in the future outcome.

(Hope that is clear, it’s late and I’m outgunned!)
 
Last edited:
Hi Itchy, The Creche example is a false analogy. It is a single seat constituency.

Or it could be likened to a system where you choose a political party first preference, and the candidate second preference. I vote for People Before Profit and then Brid Smith ahead of Paul Murphy. You vote for Fine Gael, but you don't care that if PBP gets the single seat, which of them gets it.


But it's an interesting idea, so let me make it equivalent to the Irish system for transferring surpluses and see if it helps.

Our estate of is asked how do we want the management budget spent?
A : a creche
B: a church
C: a gym
D: a coffee shop
E: a brothel

The top 3 will be picked by a Singe Transferable Vote

The 1st preferences were as follows
1,000 picked a creche,
600 picked a gym
580 picked a church

The creche option needs only 800 votes. So there is a surplus of 200 to be distributed.
Of the 1,000 who picked the creche,
300 picked a church
200 picked a gym
500 had no preference

Under the Scottish system, the TV would be 200/1000 = .2
The church would get 60 votes
The Gym would get 40 votes
So the church gets 20 more.

Unser the Irish system the TV would be 200/500 = .4
The church would get 120
The gym would get 80
So the church gets 40 more.

That does not matter to those who did not express a preference.
But it matters to those who picked a gym in the first place. The church has now gone ahead.
 
Last edited:
The 300 who picked a church as their second preference has seen their vote uses as follows:

0.8 of their vote was used to select the creche
0.4 of their vote was used to select a church.

So they have effectively had 1.2 votes

Under the Scottish system
0.8 of their vote would have been used to select a creche
0.2 of their vote would have been used to select a church.
So the total would be 1 vote.
 
If 100 people had selected a brothel as first choice

And their second preferences were
30 picked church
20 picked gym
50 made no second choice.

Presumably, you would give the church only 30 votes and not 60?
 
Now change the choices to

A. Catholic Church
B Catholic School
C Protestant School
D Muslim School
E Non-denominational school

The Catholic Church will get an advantage as their voters are much more likely to transfer from the Catholic Church to the Catholic School.
 

I think this is why you are having difficulty in seeing it. You are thinking in terms of ballot papers rather than fractions of votes.

You do have 400 physical ballot papers with no. 2 on them, but 0.8 of each one has already been used. So you really have only 80 available ballot papers.

Brendan
 

As I have shown in the church example, the voter who expressed no second preference is not being unfairly treated.

But the voter who voted for the gym is as the church is getting more votes.

Brendan
 
Clever analogy and counter analogy, guys. I'm with the Boss that the analogy is not valid, which is often why analogies are used as the primary argument doesn't quite hold. (Boss the counter analogy would have more bite if the brothel got ahead of the church ) The analogy has the false comparison that the subsequent decision has no impact on half the estate. It is assumed that because a voter does not indicate a further preference they are indifferent or unaffected by the subsequent process. One can see a sort of moral rationale that goes - "if you don't vote down the card don't moan if Loola gets subsequently elected".
Not a logical "absurdity", a physical impossibility. Totally sound to gear up votes in a computerised system. In a physical system, as we have, it is possible to gear up the surplus and that's what we do but is is completely inconsistent with the treatment of eliminations.
 
Last edited:
Brendan Burgess said:
Hi Itchy, The Creche example is a false analogy. It is a single seat constituency.
Click to expand...

It was more just to highlight, that's its a valid approach to consider the views of those who care without reference to the views of those who are indifferent.


I understand your point completely. I think your difficulty is that you're thinking in terms of fractional votes rather than ballot papers



Consider this. Instead of a secret ballot we have a rolling count of the results. Each person casts their vote in order, knowing the running total.

At a certain point, say 1980 votes have been cast and the quota for the creche has been met. We know that there are 200 votes left, we know that 500 votes have not expressed a second preference. As a result, none of the remaining 200 would vote for the creche (even though they all want one, no. 1) they are going to vote for the church/gym instead because we are already getting the creche.

Now in a secret ballot, we don't know the quota has been reached (so we put the creche as our no. 1).

In the Irish system, each of those 200 remaining votes will continue into the church/gym pile as part of the surplus, same as it would if it was an open ballot.
In the Scottish system, even though were already at the quota, because 500 people are happy with either a church or a gym, 100 of the remaining 200 votes must go in the bin (despite those people having a clear preference for one or the other). Knowing that half their vote would be wasted, the voting pattern of the remaining 200 would change in the open system.
 
Last edited:
@Itchy in your original analogy, it doesn't matter whether those who have kids are given 1 vote each per the Irish system or 0.5 votes per the Scottish system, or 100 votes each to make them feel important - it won't effect the final outcome and the folk who don't have kids don't care a jot.
In an election we might make the assumption that those who do not express a preference don't care a jot (or don't deserve to care a jot) about whether those who did express a preference have their votes count (effectively) as 0.5, 1 or 100 going forward, but it can make a difference to the final outcome.
Your amended analogy fundamentally changes the rules of the game. I'll have to think about it.
 
Last edited:
Corrected following mistake pointed out by BB.
@Itchy your amended analogy has indeed given me pause. One is reminded that multi candidate parties talk of tactical voting, the best outcome presumably being to split the vote evenly. So consider the following example.
SF has 2 quotas worth of votes which are solidly transferable. There is also 1 quota for SF first prefs which have no preference.
Scenario 1: SF(1) gets all 3 first preference quotas with a surplus of 2 quotas. Transferable votes are 2 quotas for SF(2). Under the Scottish system SF(2) will pick up 2/3rds of the surplus i.e. 4/3rds quotas; under the Irish system SF(2) will pick up the whole 2 quotas surplus.
Interestingly if the quota not transferable to SF had been transferred to others in this scenario both systems would agree.

Scenario 2: the 3 quotas split evenly, then both SF candidates are elected and there is no transfer. But let's say they nearly split evenly then SF(1) will be elected with a (slightly less) 2 quota surplus and a (slightly more) 1 quota transfers. Under the Irish system SF(2) will pick up all of these transfers (but not the whole surplus) bringing them in total to 2 quotas. Under the Scottish system SF(2) would pick up 2/3rd of these transfers bringing SF(2) to 5/3rds quotas.

Not sure why multi-candidate parties have a tactical aim. If the candidates, as distinct from the party, are partly the drawing factor I can understand why they might try to boost the less familiar candidate but that is not to game the PR/STV system but to increase their overall vote.

In summary - I am still not seeing any justification for the Irish system.
 
Last edited:
You must log in or register to reply here.
Menu
Log in
Register