The Alternative Vote System in practice.

A couple of months ago, there was an article in City AM about the upcoming - and since cancelled - London Mayoral elections, which reminded me that the election uses a Supplementary Vote system (fairly similar to Alternative Vote).

From Wiki:

The election used a supplementary vote system, in which voters express a first and a second preference of candidates.

* If a candidate receives over 50% of the first preference vote the candidate wins.
* If no candidate receives an overall majority, i.e., over 50% of first preference votes, the top two candidates proceed to a second round and all other candidates are eliminated.
* The first preference votes for the remaining two candidates stand in the final count.
* Voters' ballots whose first and second preference candidates are eliminated are discarded.
* Voters whose first preference candidates have been eliminated and whose second preference candidate is in the top two have their second preference votes added to the count.
* This means that the winning candidate has the support of a majority of voters who expressed a preference among the top two.

As it happens, it made no difference to the final outcome, Khan won more votes than Goldsmith in the first round and his winning margin was higher in the second round. In the end, only ten per cent of all votes cast were re-allocated.

Interestingly, only about 15% of voters did not bother giving a second preference vote, meaning that 85% did - but in the 2011 Alternative Vote Referendum (which proposed a very similar system), only 32% voted in favour of it. People really are strange - they are stupid enough to vote against something which in practice, they actually quite like.

It also completely puts paid to the project fear scare story that the AV system would lead to candidates from extremist parties being elected.

A) So what if it does, it's a democracy. And how entertaining would it be if we had a couple of dozen Green/Socialist Workers' party MPs and a couple of dozen UKIP/Brexit Party/BNP MPs on the Opposition benches going at each other hammer and tongs?

B) Project Fear also claimed that AV would lead to more coalitions, which sort of cancels out the first claim, as coalitions tend to be more  moderate.

C) As we see in practice, it makes very little difference. The winning candidate was always going to be from one of the Big Two parties.

IMHO, AV is still a good system. It doesn't change the outcomes of elections very much, if at all - what it does change is what sort of policies the winning candidates actually implement afterwards. The only way to 'send them a message' as to what you actually want is to vote for a smaller party with a clear manifesto or a single-issue party. The AV system clearly encourages people to give their first vote to a smaller party and their second vote to one of the Big Two as a fall back.

Fun with numbers - splitting the Remain vote at the MEP elections

The MEP elections in Great Britain, to be next held on 23 May 2019, use the d'Hondt system for allocating seats in each of eleven constituencies/regions.

Whether the Remain parties (Lib Dem, Green and Change UK) have shot themselves in the foot (feet?) by competing over the same small pool of voters is an interesting question.

Let's treat this as an unofficial In-Out Referendum and assume votes cast are in line with current opinion polls and are the same in each constituency, as follows:

Leave
Brexit Party - 31%
UKIP - 4%

Remain
Lib Dem - 10%
Green Party - 10%
Change UK - 10%

Undecided - neutral - ambivalent
Labour - 22%
Tories - 13%

The more seats there are in a constituency, the closer the result is to proportional representation; the fewer seats, the closer the results are to FPTP.

If you crunch the numbers (or use Paul Lockett's fine calculator) for the largest constituency with ten seats (South East), the end result is the same whether the Remain parties had put up a single list or not - Leave 4 seats, Remain 3 seats and Undecided 3 seats.

The difference is that with a single list and 30% of the vote, Remain would win seats 2, 5 and 9; with three competing Remain parties, they will win seats 7, 8 and 9. So they will do relatively worse in smaller constituencies and relatively worse overall.

The reverse is true for Leave, only not as markedly. If Remain had put up a single list, Leave would win seats 1, 4, 8 and 10 of a ten-seat constituency. With the Remain vote split, they will win seats 1, 3, 5 and 10.
-----------------------------------------------
To sum up, for various sizes of constituency with a split Remain vote, seats will be as follows:

3 seats = Leave 2, Undecided 1
4 seats = leave 2, Undecided 2
5 seats = Leave 3, Undecided 2
6 seats = Leave 3, Undecided 3
7 seats = Leave 3, Undecided 3, Remain 1
8 seats = Leave 3, Undecided 3, Remain 2
9 sweats = Leave 3, Undecided 3, Remain 3
10 seats = Leave 4, Undecided 3, Remain 3

As only five constituencies have seven or more seats, Remain have definitely messed up badly. With a single list Remain vote, seats would be as follows:

3 seats = Leave 1, Undecided 1, Remain 1
4 seats = leave 2, Undecided 1, Remain 1
5 seats = Leave 2, Undecided 1, Remain 2
6 seats = Leave 2, Undecided 2, Remain 2
7 seats = Leave 2, Undecided 3, Remain 2
8 seats = Leave 3, Undecided 3, Remain 2
9 sweats = Leave 3, Undecided 3, Remain 3
10 seats = Leave 4, Undecided 3, Remain 3

Just sayin'...

Fun with numbers/FPTP

From the BBC:

According to analysis by polling expert John Curtice, if the results of Thursday's polls in Wales, Scotland and 32 county councils in England were repeated nationally, the Conservatives would be on 38%, Labour 27%, the Lib Dems 18% and UKIP 5%.

With some form of PR, the best being probably multi-member constituencies, there'd probably be a slight left/centre-left majority in the House of Commons, which is the opposite of the slight Leave majority in last year's Referendum.

Not sure who'd get the missing 12% of the votes, presumably the regional/nationalist parties, SNP, Plaid Cymru and so on, who tend to be (seen as) more on the left.

Fun With Numbers

US election, share of popular vote:

Hillary Clinton, Democratic Party, 48%, 59,814,018 votes

Donald Trump, Republican Party, 47%, 59,611,678 votes

Gary Johnson, Libertarian Party, 3%, 4,058,500 votes

Jill Stein, Green Party, 1%, 1,213,103 votes

This is how George W Bush was elected in 2000, as it happens.

And well done, Gary Johnson, I would have voted for you*, and failing that I'd have voted Jill Stein just for the heck of it.

* Obvs, he's a Faux Libertarian not a proper libertarian, as the American Libertarian Party thinks that VAT is better than income tax and income tax is better than LVT i.e. they have it completely arse-backwards, but hey.

Interesting... even if possibly a little biased.

From The Electoral Reform Society:

Councils dominated by single parties could be wasting as much as £2.6bn a year through a lack of scrutiny of their procurement processes, according to a new report for us released today.

The study – undertaken by Cambridge University academic Mihály Fazekas – is titled The Cost of One-Party Councils and looks at the savings in contracting between councils dominated by a single party (or with a significant number of uncontested seats), and more competitive councils...

The report also measures councils’ procurement process against a ‘Corruption Risk Index’ - and finds that one-party councils are around 50% more at risk of corruption than politically competitive councils. The corruption risk of competitive councils compared to those dominated by one party is similar to the difference between the average Swedish municipality and the average Estonian municipality. This doesn’t bode well for democracy or council coffers.

And it’s no small-scale study. It uses ‘big data’ to look at 132,000 public procurement contracts between 2009 and 2013 to identify ‘red flags’ for corruption, such as where only a single bid is submitted or there is a shortened length of time between advertising the bid and the submission deadline.

In case you were wondering how they calculate the number of MEPs each party gets in each region...

… I refer you to Paul Lockett's fine calculator here.

He set it up back in 2009, but the system doesn't seem to have changed and it still works.

Fun Online Polls: Lost Causes and EU propaganda

Thanks to everybody who took part in last week's Fun Online Poll, excellent turnout of 169 in only four days. I deliberately didn't have a 'none of the above' option because I was depressed enough. Results as follows:

Which of the following "lost causes" do you support?

Above the line (more than 85 votes):
Allow smoking in pubs and cafes 122 votes
Leave the EU, the UN etc. 121 votes
Legalise cannabis 111 votes
Legalise brothels 108 votes
Allow fox hunting 105
Replace entire welfare system with a Citizen's Income 93 votes

Below the line (fewer than 85 votes):
Replace old age pensions with a Citizen's Pension 78 votes
Replace as many taxes as possible with Land Value Tax 75 votes
Use Proportional Representation 69 votes
Turn off the traffic lights 59 votes

The Citizen's Pension was the joker in the pack as Iain Duncan Smith is going to do it anyway; I'm obviously disappointed about lack of support for LVT; proportional representation is a secondary issue; and I'm particularly surprised about the lack of enthusiasm for turning off traffic lights.
----------------------------------------
And lo, to this week's Fun Online Poll.

We've established on several occasions that the readers of this 'blog would vote to leave the EU in an In-Out referendum, and it's just about conceivable that if there were a snap In-Out referendum next week that a slim majority would vote for "Out" but that's not really the issue.

As we have learned from the No2AV campaign, you can get any result you want if you pump out enough propaganda; and I imagine that if They announced that there'd be In-Out Referendum in six months' time, then we would be subjected to a 24/7 barrage of a few simple - but somehow plausible - lies for those six months and the result would be "Stay in".

So this week's Fun Online Poll is whether you think that They would be able to pump out enough propaganda to swing the vote their way.

Vote here or use the widget in the sidebar.

Fun Online Polls: AV and Lost Causes

Thanks to everybody (nearly two hundred votes in four days!) who cast a vote in this week's Fun Online Poll, which I might as well shut down now that polling has closed. Results as follows:

How will you vote in Thursday's referendum?

Yes to AV - 51%
No to AV - 37%
I won't bother - 12%

I doubt whether the official result will be as favourable. Ah well.

This does not bode well for an In-Out Referendum on the EU. All They need to do is choose a few simple lies and stick to them, such as Three million jobs depend on our membership of the European Union (scroll down to the end of Channel4's Fact Check service) and that will be the end of that for a few decades
------------------------------------------
Which brings me back to my favourite topic of "Lost causes". It seems that pretty much anything I campaign for (actively or otherwise) is pretty much a lost cause, so out of interest, how many of my favourite "Lost causes" do you support?

Vote here or use the widget in the sidebar, multiple answers allowed.

"FPTP is unpopular"

On page 5 of their leaflet, the No to AV people claim that "AV is unpopular" because only three countries use it.

What they don't tell you of course is that "FPTP is [only] the second most widely used voting system in the world, after party lists. In crude terms, it is used in places that are, or once were, British colonies. Like America. The use of FPTP used to be even more widespread, but many countries that used to use it have since switched to something else."

So yes, FPTP has been "tried and tested" as the No to AVers say... and it has been found wanting. If we were going by "popularity" alone, we'd go for full PR and party lists, and there'd be no argument for going back to FPTP.

And maybe AV isn't the best voting system (although it's clearly much better than FPTP) but wasn't it the existing government that offered us the choice? If they think AV is so awful, why didn't they offer us a choice between FPTP and (say) FPTP-with-top-up-seats?

"I'm saying yes!"

Pro-FPTP deviousness

I have no idea how or why the Conservatives and Lib Dems agreed that there'd be a referendum on changing from FPTP to AV or which alternative voting system the Lib-Dems really wanted (which is not necessarily what they say they wanted, of course).

Nonetheless, we are to be offered a choice between FPTP and AV.

The No2AV crowd, which is basically the Conservatives and a few Labour MP dinosaurs who support FPTP, are now trying to discourage people from voting for AV on the basis that "only three other countries in the world use it" (which turns out not to be true either), and by implication that it isn't very good.

That is how devious the Conservatives are; to offer us a choice between something that is provably shit, like FPTP (which only a minority of countries use) and something which by their own admission isn't much better. So why didn't they give us a choice between FPTP and one of the other versions of Proportional Representation which is much better than AV*?

* I think multi-member constituencies is the best system, others may disagree. I'd have happily gone with whatever the Lib Dems wanted, seeing as they're in government and not me.

Hilarious AV versus FPTP debate yesterday on Channel 4 News

Spotter's badge: Neil Craig.

George Osborne doesn't do logic either

Or facts. From The Daily Mail:

Mr Osborne said: ‘What really stinks is actually one of the ways the Yes campaign is funded. The Electoral Reform Society, which is actually running some of the referendum ballots, and is being paid to do that by the taxpayer, stands to benefit if AV comes in... that organisation, the Electoral Reform Society – part of it is a company that makes money – is funding the Yes campaign.

'That stinks frankly and is exactly the sort of dodgy, behind the scenes shenanigans that people don’t like about politics. The No campaign has asked for it to be investigated by the Electoral Commission and certainly I think there are some very, very serious questions that have to be answered.’

So what?

The ERS has been campaigning for PR (and by default AV over FPTP) for ages, and it has a subsidiary which prints ballot slips. There's no law says that contracts for printing ballot slips in future have to go to their subsidiary, and the idea that AV needs 'expensive counting machines' is nonsense.

You can do the numbers yourself, but by and large the number of ballot papers which have to be shuffled from one pile to another is about one-fifth of the votes cast, so that's the difference between counting a two-thirds turn-out and counting an eighty per cent turn-out. I've had run-ins with El Comm, what they really hate is donations from undisclosed parties; the Yes2AV have never made a secret about who gives them money... unlike the No2AV people, of course.

He claimed AV could have distorted key election results over previous decades as people’s second, third and subsequent preferences altered the outcome and allowed parties to stitch up coalition deals behind closed doors.

Yes, AV might have led to slightly different results than FPTP; that's the whole point. We could equally argue that FPTP produced a 'distorted' result compared to what the result under AV would have been.

As to 'stitch ups behind closed doors', good grief, try looking in the mirror, Georgie! And we will be able to see the electoral price that the Lib Dems have to pay for this in a few weeks' time - these 'stitch ups' are not a no-risk strategy.

Georgie then keeps on digging:

'For many people in this country their vote won’t count as much as voters who vote for extremist parties. If you want a fair voting system – if you want your voice heard in this country and want a say over how your community is governed, you would want your vote to count at least as much, if not more, than a BNP voter.'

Under FPTP votes are far from equal; the vote of a swing voter in a marginal constituency is worth infinitely more than a vote in a safe seat. Be that as it may; the idea that a ballot slip saying e.g. "BNP -1; Labour 2" has more influence on the outcome than one just saying "Labour - 1" is palpable nonsense; by and large they both have the same impact.

In this example, if the Labour candidate wins on the first round, the BNP/Labour ballot slip has no weight at all; if Labour wins after the BNP/Labour ballot slips are redistributed then the Labour and BNP/Labour ballot slips have the same weight; and if the Labour candidate doesn't win anyway, then neither the Labour nor the BNP/Labour ballot slip has any weight at all.

As ever, let me point out that the Tories are daft buggers: the single-member FPTP system is biased against them (for obscure mathematical reasons); what they should have proposed a referendum on is 'Multi-Member Constituencies' which is even better than AV, so there is no way the Lib Dems could have turned this down at the coalition negotiations.

David Cameron doesn't do logic...

From The Daily Mail:

The good thing is that when people hear the arguments [against the Alternative Vote system], they do come down on the No side. If we can get people to focus briefly on the arguments – people who vote for extremist parties have their vote counted more times (1), candidates who come third can end up winning, it’s a more complicated system (2), it’s more expensive (3), it isn’t necessarily more fair (4) – I think we can win.’

1) Simply not true. If somebody just casts their first vote for (say) Labour on a ballot paper with five candidates, then that is exactly the same as casting their first, second, third, fourth and fifth votes for Labour. Their ballot slip will be counted (or taken into account) just as many times as somebody else's ballot slip who votes (say) BNP first, Labour second. And a ballot slip with "Tory 1" is counted just as many times as a ballot slip with "UKIP 1, Tory 2", and so on.

And the kicker is this: if somebody just votes "BNP 1" and leaves it at that, the chances are that his ballot paper will only be counted once, so on closer inspection, the situation is anything the opposite of what Cameron says.

[As an aside: if these pro-FPTP politicians had any principles whatsoever, they would announce that they would refuse to accept their seat if the only reason they won the election because more ballot papers showing a vote for the BNP (or another 'extremist' party - presumably they mean UKIP? The Greens? Plaid? SNP?) than the second candidate.]

2) No it's not. You put a "1" against your favourite candidate, a "2" against your second-favourite and so on. This won't take any longer than standing in the polling booth thinking "OMG, who is the least-worst out of all these numpties?"; it might even be quicker.

3) Agreed, it will make the actual counting procedure take about a tenth longer, but the counting itself is only a small part of the cost of running an election and this is a price worth paying.

4) FPTP "isn’t necessarily more fair" either, the point is to allow people to cast more 'protest' votes. Yes, AV might lead to some quirky results, but so does FPTP, across six hundred constituencies, it will all even out.

And the intellectual giant managed to completely contradict himself within the space of one speech:

Exhibit A: "He said that once people realise AV... boosts the influence of those backing the BNP and other extremist parties... they back the No campaign.

Exhibit B: "The Prime Minister also warned that AV would lead to more tentative and boring politicians, as candidates would be careful not to offend anyone because they would be concerned about the second preferences of voters."

Is it just me or are those two statements completely at odds?

FPTP is shit, even by the admission of supporters of FPTP

Deniro, in the comments:

The counting system used in AV does not work. It incorrectly gives special significance* to the second choice votes of voters who vote for lower ranking first vote candidates.

The second choice votes of voters who vote for higher ranking first choice candidates are just as significant in ascertaining which candidate has a majority endorsement. Therefore it declares a winner incorrectly. The AV count system simply does not work.

Doesn't that apply in spades to FPTP? To paraphrase:

The counting system used in FPTP does not work. It incorrectly gives special significance to the second choice of voters who did not dare vote for their preferred candidate.

The second choice votes of voters who vote tactically are just as significant in ascertaining which candidate has a majority endorsement as those who genuinely prefer that candidate. Therefore it declares a winner incorrectly. The FPTP count system simply does not work.

* It does not give special significance to second pref! That is a myth!

Imagine somebody's popping to the corner shop and they ask you if you want anything. You tell him "I'd like a bottle of coke, but if they're sold out, bring me an orange juice." He comes back with an orange juice because coke was sold out. Has your second prefence been 'special significance'?

No2WS

From Mark Reckons via Neil Harding.

Genius

James Ward decided to email the No2AV campaign:

Hi,

A couple of days ago, an advert for the No2AV campaign appeared in the Birmingham Mail featuring a picture of a baby girl and the message “She needs a new cardiac facility NOT an alternative voting system”.

The advert explained that the £250million which would be spent switching to AV could be better spent on a new Children’s Heart Centre at Birmingham Children’s Hospital. I was just wondering if you had any background information on who the baby is and what her condition is. I really don’t want the baby to die just because I voted “yes” and so I want to have all the facts at hand before I decide how to vote.

This is the deciding factor in whether I vote “yes” or “no”, so any information would be very much appreciated.

Thanks,
James

I hope they reply soon, my vote hangs in the balance.

AV: Fun With Numbers

The FPTP gang claim that the counting procedures would be prohibitively expensive under AV, which intuitively is complete bollocks (the real cost is the people updating the electoral register, printing the ballot papers, manning the polling stations, ferrying around the sealed boxes to the final count etc, all the TV and press coverage etc, not the last bit of paper shuffling).

I trust that even the most hardened FPTP supporter is prepared to accept that the system should be geared up to bearing the cost of a 100% turn out? Seeing as there's usually only a 60% or 70% turnout, then as long as the number of votes recounted (however many times) is not too high, the total cost of the initial count and the redistributing would still be less than the cost of counting all the votes were there to be a 100% turnout under FPTP.

OK, so let's assume that at the next General Election, held under AV, people vote in similar proportions to the 2010 General Election; that the median number of preferences indicated by any voter is 2 (so at each elimination stage, half of ballot papers of the lowest-scoring candidate are recounted and half are discarded); and that there is no particular correlation between how people cast their first, second and subsequent preferences.

The simple fact of the matter is that the number of times a ballot paper is picked up and put on a pile is only 28% more than under FPTP, see if you can follow the maths:The nice bit is that instead of 74% of votes being 'wasted', only 62% are 'wasted' (which must be an improvement).

"Aha!" shouts the crowd, "What if people are suitably emboldened at the next election and first preferences are spread far more evenly?"

Doesn't make a big difference. Let's now assume the BNP gets 10% of first preferences and the other five candidates get 18% each, the total number of times a ballot paper is sorted onto a pile is still only 37% more than under FPTP. In this example (assuming the Tory candidate would have edged in) only 68% of votes are 'wasted', which is still an improvement on the 76% that would be 'wasted' under FPTP.To sum up:

Will AV increase the cost of running elections? Nope, not materially.

Will AV change the outcome (in terms of which parties get how many MPs)? Probably not much.

Will AV embolden the smaller parties and add to the gaiety of the nation? Yes of course.

Will AV remind MPs that none of them represents anywhere near a majority of voters? Yes - and that is surely the whole point.

Sun readers thick: official

From The Soaraway Sun*:

The 'No' vote last night soared to a seven-point lead in a Sun YouGov survey. Forty-one per cent said they want to stick to the current 'first past the post system', while 34 per cent want to change it to the complicated Alternative Vote.

So semi-literate Sun readers can manage to find their favoured candidate on a ballot paper and mark an "x", but finding their favoured candidate and marking it with "1" is beyond them? Maybe even the few borderline literate-numerate Sun readers might even manage to mark their favoured candidate with a "1" and then mark their second choice with "2"?

As Neil Harding explains:

If you ever said to someone going to the shops - "Get me a coke or if they haven't got that, I'll have a lemonade", then you understand the principle behind AV voting. It only sounds complicated if you explain it badly, which the No campaigners are doing on purpose (finding the most wordy academic text they can).

He duly extends the analogy**:

If you order a chicken curry at a restaurant, but are told that has sold out then decide to have a lasagne instead, you have only had one meal. The same is true for AV, only ONE of your preferences will count towards the end result. Don't be fooled by propaganda saying otherwise.

* Spotted by Denis Cooper, who's firmly in the AV camp.

** For the benefit of Sun readers, an "analogy" is a simple example used to explain something, and has nothing to do with the study of people's bottoms.
-------------------------------
PS, if you have ever attended a count, you'll know that the extra work involved with AV would be fairly minimal. Under FPTP, tellers make a pile of ballot slips for each candidate (in bundles of twenty or something) and then the biggest pile wins (they count them again, under the eyes of the candidates, if it looks fairly close).

The same basic system would apply under AV, only if no candidate gets more than half the first choice votes (which will happen in most constituencies), they'll just grab the smallest pile and redistribute it; and then the next smallest pile and so on. Mathematically, it's unlikely that more than the thirty or forty per cent of the ballot papers would have to be picked up more than once or twice, and as there will only be a few dozen or a couple of hundred in the smallest piles, that's no big deal.

Apart from a few dedicted anarcho-democrats who rank all candidates in reverse order of how likely they are to be elected, I doubt sorely whether most people will use more than their first and second votes, which reduces the amount of re-allocating even further.
-------------------------------
PPS, I can only assume that Rupert Murdoch (who owns The Sun newspaper) has established that he's happy 'doing business with' a Labour government or a Tory-led government, but doesn't have the time or inclination to have to 'do business with' (i.e. bully and bribe) a load of smaller parties as well.

Fun Online Polls: Votes for prisoners, Wills and Kate

Thanks to everybody who took part in last week's Fun Online Poll (a high turnout with 149 votes, but it ran for two days longer than normal). Results as follows:

What do you think of 'votes for prisoners'?

All the more reason to leave the EU (even though this was an ECHR decision) - 41%
I am opposed - 37%
I am not bothered either way - 15%
Other, please specify - 7%

So there we have it. I personally am not bothered either way, but if the issue helps stoke anti-EU sentiment, even if the case was decided by the European Court of Human Rights, which is a quite separate institution, then it's all good.

As far as I can see, as long as we have first-past-the-post, then most people's votes are effectively worthless anyway, so whether prisoners are allowed to cast a worthless vote or not* is neither here nor. Indeed, I lived in Germany for nine years and being a foreigner, wasn't allowed to vote at all, big deal.

* As Leg-Iron points out, it might get a bit icky if all inmates were allowed to vote in the local elections wherever the prison happened to be.
---------------------------------
On a lighter note, apparently a celebrity jobless couple from the South of England, referred to as Wills and Kate by the media have decided to get married at some unspecified point in the future. Our current Prime Minister, possibly the most out-of-touch politician in British history, responded thusly:

"As well as this being a great moment for national celebration, I think we also have to remember that this is two young people who love each other who have made this announcement, who are looking forward to their wedding, and we must give them plenty of space to think about the future and what they are about to do."

How about giving a moment's thought to all the young couples who have jobs and are thinking of 'settling down', and addressing the issue of why they can't afford to buy a house, let alone for Mum to give up work to have kids (unless they are out of work, in which case they would appear to have our Prime Minister's blessing), eh?

So that's this week's (belated) Fun Online Poll: "Wills and Kate have decided to get married: what do you think?"

Vote here or use the widget in the sidebar.