UKIP as a spoiling factor, part II

Following on from yesterday’s post on whether or not the presence of UKIP had cost the Conservative party a number of seats…

cim has done some analysis suggesting that a UKIP candidate produces a net increase in turnout of about 2.5% – 1,000 votes. However, the average UKIP candidate took around 1,600 votes, suggesting that they’re only taking around 40% of their votes from other candidates and the remainder from potential abstainees. If about half of that 40% would have voted Conservative, and the remainder voted for another minor party if one was present, then we can find the seats where the Conservative candidate was less than 20% of the UKIP vote away from beating their opponent.

There’s six of them – albeit one by a mere four votes, so we can perhaps call Dudley North a 50-50 chance.


If we raise the defection rate to 30%, we still have the same six seats under the threshold; if it’s raised to 40% – indicating that all defectors to UKIP came from the Conservatives – they would recover a seventh seat, Great Grimsby, from Labour.

Six seats is not enough to produce a definite majority, of course – but it’s certainly a non-trivial number, especially since they would come directly from the opposition. Had they taken these six seats, we’d be looking at 311 for the Conservatives versus 301 for an opposing Lib Dem-Labour coalition; it would become substantially harder for a reasonably robust coalition to form against the Conservatives, and so they would have a better chance at gambling on a minority government.

Historically speaking, it’s a plausible number. The Referendum Party (remember them?) in 1997 took 2.6% of the votes, and are estimated to have cost the Tories about four seats. UKIP took 3.1% of the vote this time around, so six seats is well within the bounds of plausibility.

Proportionalising the results

I was curious to see how this election would have played out under a proportional representation system. We can’t easily guess what an STV result would have looked like, since it would involve very different voting dynamics – but we can make a rough guess at what would happen if we had used multi-member regional seats, similar to the regional section of the system in Scotland.

(Scotland, for those unfamiliar, has a mixed system – every person is in an individual constituency, where they cast a normal FPTP vote; they then cast a second vote to elect members from their region, which returns about half a dozen MSPs. Regional votes are counted in a moderately complicated way so as to benefit parties with lower local representation; it actually comes out pretty representative of overall vote-share.)

Let’s say we do away with local seats entirely, and implement simple voting in county-based seats, with as many members from that county as it gets under the current system. Because people only cast a single vote, their preferences are likely to be broadly similar to their preferences under FPTP, and so we can use the existing data from the past two elections… and why counties? Constituencies are already grouped to follow county boundaries, and it provides a reasonably natural way of dividing them up.

What would it look like? A handful of arbitrary examples (2005 results are using 2010 boundaries; seat allocation uses the nice simple D’Hondt method):


  • FPTP 2005 – 6 Con, 1 LD
  • PR 2005 – 3 Con, 2 Lab, 2 LD
  • FPTP 2010 – 6 Con, 1 LD
  • PR 2010 – 4 Con, 1 Lab, 2 LD


  • FPTP 2005 – 1 Con, 13 Lab, 1 LD
  • PR 2005 – 3 Con, 9 Lab, 3 LD
  • FPTP 2010 – 1 Con, 13 Lab, 1 LD
  • PR 2010 – 3 Con, 9 Lab, 3 LD


  • FPTP 2005 – 4 Con, 1 Lab, 1 LD
  • PR 2005 – 3 Con, 1 Lab, 2 LD
  • FPTP 2010 – 5 Con, 1 Lab
  • PR 2010 – 3 Con, 1 Lab, 2 LD


  • FPTP 2005 – 3 Con, 3 Lab
  • PR 2005 – 3 Con, 2 Lab, 1 LD
  • FPTP 2010 – 6 Con
  • PR 2010 – 3 Con, 2 Lab, 1 LD

That last one is perhaps the most interesting – under FPTP it swung wildly from 3-3 to 6-0, but under PR it remained a balanced 3-2-1.

Some open issues with this model, though, frivolous as it is –

a) To what extent are the figures we’re getting from FPTP data representative of tactical voting, not of true preferences? That would drive down the results of third & fourth parties in any given region, and boost the second-placed.

b) What’s the optimal size for a region, and how do we allocate them? My gut feeling is that around half a dozen members is approximately right – big enough to iron out local oddities, small enough not to be too unwieldy – but I’m not sure if this is based on anything in particular. (Seven-member seats mean that each has approximately 700,000 people in it – the same size as the average US congressional district)

c) Shouldn’t we be using STV in these regions anyway? (The answer to that one is: probably yes. But see above.)

UKIP: defeating the Conservatives?

Here’s a thought.

The Conservatives need 326 seats for a majority. They took 306; they get another eight from the support of the DUP, putting them on 314 and needing a mere twelve seats to govern. They almost certainly could have counted on the votes of any parties to the right of them which got elected – at least, not too far to the right – which, in practice, means UKIP. UKIP, of course, did not get any MPs.

But how many did they cost the Conservatives? In effect, had UKIP not run, the bulk of their voters could have returned to the Tories – would this have provided Cameron with any extra seats?

The answer, interestingly, is quite a few. There are 22 seats where the Labour MP’s majority is less than the number of votes polled by UKIP, and seven where the Lib Dem MP’s majority is less than the number of votes polled by UKIP. Of those, fifteen of the Labour and and six of the Lib Dem seats were marginals with the Conservatives – potentially, UKIP cost the Conservatives up to twenty-one seats, nine more than the critical number. and enough to govern (just) without needing the DUP.

Now, not all those voters would have returned to the fold. Some would have defected to the BNP – though it did not run in all those constituencies – and those primarily intent on casting an anti-establishment vote would have defected to the other minor parties or stayed at home. Let’s assume that half the UKIP voters defected and the other half cast a mixed ballot of minor parties or abstentions.

In this case, they would have taken three Lib Dem seats and nine Labour seats; enough to either put the Conservatives in power or in a minority position so near to it that a coalition to bring them down would be unworkable.

It’s a surprising result, I have to say – but if we end up with a Con-Lib coalition, a government of the Conservatives essentially tied down from doing anything too stupid by being dependent on a more moderate party for support… we might well have Nigel Farage and his friends to thank for the difference between this and a Conservative government ruling unhindered.

[See here for Part II]

General Election: everyone lost

It has been a hectic few days. But let’s quickly recap the results:

Labour: Well, they lost, and substantially more so than in the other close election, 1992. Net loss of ninety seats; fifty seats behind the Conservatives and a vote share of under 30%, the worst support from the population since the heyday of electoral suicide notes in 1983.

Conservatives: The Conservatives had a following wind from the media. They had a financial crisis and a corruption scandal denting the reliability of the governing party; they were up against a personally unpopular prime minister; they had a new, charismatic – well, by Tory standards charismatic – leader and a party that had had almost fifteen years to recharge itself in opposition. And they failed to get to a majority, partly because they were still recovering from the utter catastrophe of 1997 and partly because the party simply still isn’t trusted by the country.

Liberal Democrats: competing against two rather inept main parties, they improved vote share by a mere 1%, and actually lost a number of seats – something that hasn’t happened for a long time.

The minor parties had an at best mixed night:

UKIP: 3% vote share, but lost 457 (!) deposits. Failed to gain any MPs, and – let us not forget – came within a whisker of their leader dying on polling day.

BNP: Two percent of the vote, but losing 80% of their deposits, and nowhere near the results needed to get any seats. The absolute number of voters was half that in the 2009 election – suggesting the core of nutters is strongly outweighed by the protest-vote element.

Greens: on the one hand, an MP! On the other hand, they lost 98% of their deposits, and actually polled fewer votes than last year.

SNP: stalled. No improvement on seat numbers; despite the potential benefits of status coming from local incumbency and the (perhaps vindicated) argument they could be key in a hung parliament, they didn’t pick up significant numbers of votes.

Plaid: lost vote share, no gains. I admit I do not know enough about the Welsh background to say more than this – but they lost a substantial share of the vote in the close marginal of Ceredigion, and fell back in Ynys Mon, both seats they must surely have hoped to take.

The one party who unquestioningly did do well was the Alliance in Belfast. Against that, Sinn Fein and the SDLP both lost vote share; the DUP lost one of their strongholds (and their leader!), and the UUP lost their only seat because the incumbent split from the party and ran as an independent.

Thought for the day

The lowest vote share a party has taken in a general election and still held a majority is 38% – the Conservatives in 1923. Labour formed a minority government in 1929 on the back of 37.1% of the vote share; and a minority government again in February 1974 on the back of 37.2% of the vote. …35.3%, in 2005. Goodness, do I feel a bit silly for missing that one – thanks cim for spotting it.

It’s likely that whatever result we get on Thursday, we’re going to break that record; barring a sudden Conservative surge, a hypothetical narrow-majority government will have the lowest share of the electorate’s support any majority government has ever had. A minority Conservative government – if the Lib Dems pass up coalition – would quite possibly, again, have the lowest level of support recorded for a sole party in power.

Reassessing: on current polling, it’s likely the Conservatives will either form a majority government on slightly above the 2005 vote-share, or potentially a minority government – if a Lib-Lab coalition fails – on slightly below it. Not the best of omens for making sweeping changes, but it’s remarkable to see the dramatic difference – in seat numbers and in general perceptions of strength – between what the Conservatives would get with 36% tomorrow, and what Labout got with 36% half a decade ago.

Signalling changes

From the Guardian, musing on the handy symbolism of coalition colours:

It’s maybe worth noting that, in the international nautical code, a striped blue-and-yellow flag means “I require a pilot”; two red and yellow triangles, “man overboard”.

Why stop there? There’s a whole alphabet of symbolism:

  • A large blue centre with yellow lurking around the edges: “Keep clear of me; I am manoeuvering with difficulty”.
  • Equal blue and red (!): “I am altering my course to starboard”.
  • Equal yellow and blue: “I wish to communicate with you.”
  • Red cross on yellow: “The way is off my ship” – I am not moving and you may pass safely.
  • Red and yellow stripes, closely mingled: “I am dragging my anchor.”

The jokes write themselves, really – especially the first and last ones.

It doesn’t just work for coalitions, either: solid red with a swallowtail indicates “I am taking in, or discharging, or carrying dangerous goods”. There is no solid blue flag – presumably it doesn’t show up well at sea – but the various blue-and-white permutations are rather unfortunate. A blue square on a white field indicates “I am moving backwards”, whilst a white stripe over blue is “I am leaking dangerous cargo” or, indeed, “I am on fire”. White and blue checked is simply “negative”, and a blue cross on white is, appropriately, “Stop carrying out your intentions and watch for my signals”. Red surrounded by blue is, fittingly, “I require medical assistance”.

The Lib Dems come out perhaps the best of a bad bunch – solid yellow indicates that “my vessel is free of disease and I request permission to enter harbour”.

Of course, combinations of flags could mean things as well. A Lab-Lib coalition might reasonably be represented by BO or BR – red dominant over yellow – representing, respectively, “We are going to jump by parachute” and “I require a helicopter urgently”. Maybe it’s a fairer coalition – OR, a bit more evenly split, means “I have struck a mine”… and reversed, it’s “My propeller shaft is broken”.

Perhaps, on the whole, this is something that does not bear looking into too closely.

On Philippa Stroud

From the Guardian:

Last weekend The Observer revealed that Philippa Stroud, the head of a thinktank set up by former Conservative leader Iain Duncan Smith (the Centre for Social Justice), and the Conservative candidate for Sutton and Cheam, has been trying to drive demons out of lesbians, gay men and transsexuals. (…)

Question one: Why hasn’t Cameron asked for Stroud to be deselected and distanced himself from her thinktank?

Last week, the Tory leader said that he decided to suspend Philip Lardner, the Conservative candidate for North Ayrshire and Arra, “within minutes” for writing on his website that he thought homosexuality was “not normal”. Lardner also opined that “most people” consider homosexuality to be “somewhere between unfortunate and simply wrong”. A hate-soaked, erroneous diatribe, but compared to Stroud he looks like Peter Tatchell. Why the discrepancy? How can abusing young people with extremist religious practices be less incriminating than words?

I fear the answer to that one is cynically pragmatic: the Conservative candidate in Sutton and Cheam needs a mere 3,000 votes to take the seat from the Lib Dems. (Electoral Calculus give her about a one in three chance)

Meanwhile, the erstwhile Conservative candidate in Ayrshire North would have needed to take about 11,000 votes from Labour – and not lose any to the SNP, who were neck and neck with him. (Electoral Calculus gives him about a one percent chance)

Far be it from me, of course, to say that whether or not David Cameron gets distressed by bigotry has anything at all to do with whether it’s likely to cost him an extra seat…

The process of a hung parliament

The Cabinet Secretary prepared a paper earlier in the year on the transition periods between governments, including the process for a hung parliament – it’s been published here.

The key quote is: “Where an election does not result in a clear majority for a single party, the incumbent Government remains in office unless and until the Prime Minister tenders his and the Government’s resignation to the Monarch. … As long as there is significant doubt whether the Government has the confidence of the House of Commons, it would be prudent for it to observe discretion about taking significant decisions, as per the preelection period. The normal and essential business of government at all levels, however, will need to be carried out.”

In other words: whatever happens, short of a clear majority, Brown remains in power until the dust settles.

As to just how the dust settles, this summary points out an interesting wrinkle: the Lib Dems have a system in place to control how a coalition is decided upon. Clegg would need the explicit support of three-quarters of the MPs and of the party executive; failing that, he could appeal to a special conference, and need a two-thirds majority there; failing that, he could take it to a ballot of all party members. I can’t see it ever dragging on that long, but it is interesting to know that it’s not simply a matter of the party heads making a decision and then moving swiftly on to government.

Turnout figures

Liberal Conspiracy, I’ve noticed, are quoting the YouGov polling figures in full – not just the headline percentages, but some of the subsidiary questions. (14/4; 18/4).

The one that interests me is the “likelihood to vote” question. Last week, 65% of those polled said they were “absolutely certain” to vote; this week, 68%. Now, these percentages are probably overoptomistic – people do like sounding politically involved and keen when asked questions, and there are of course no shortage of people who intend to vote, but get stuck in traffic coming home from work, or come down with the flu, or simply forget.

If we assume everyone overestimates their willingness to vote by ~10%, then we’re looking at an expected turnout of 75-78%. (If we assume they overestimate by ~20%, then we’re looking at an expected turnout of 66-68% – less remarkable, but still better than recent standards)

I wonder if we have the polling likelihood question from previous years, to compare the actual result and see what our modifier should be…

(Also: it’s interesting to see the likelihood of voting ticking up along with the recent swing to the Lib Dems. Cause or effect?)