Saturday, 9 May 2015

Liberal Democrats Win General Election

The last time I felt like this, Arsenal had just been hammered 6-0 by Chelsea in Arsene Wenger's 1,000th game. This is not because I was disappointed - I had thought that we might get a draw at Stamford Bridge last year and have had a nagging suspicion for a couple of weeks that the Tories would cobble together another coalition - but because avoidable mistakes had been ruthlessly exploited. It didn't have to be this bad. One prediction I feel safe in making is that Labour will now reimpose the "democratic centralism" of the 90s, which means a managerialist (and thus Blairite) resurgence. The prominence of Alastair Campbell on election night and after, regretting the inadequacy of Labour's "narrative", is worth noting, as much as the queue of "modernisers" urging Labour to rediscover "aspiration" and "celebrate our entrepreneurs and wealth creators".

One of the great excuses of football failure is "we didn't turn up", and that appears to be a large part of the story in Labour's defeat. Liberal moralists see the result as a judgement on our sins, hence trendy-rev Giles Fraser is "Ashamed to belong to a country that has clearly identified itself as insular, self-absorbed and apparently caring so little for the most vulnerable people among us". For Polly Toynbee, the voice of apocalyptic centrism, "The future is dark. In England and Wales the people swung to the right". This is nonsense. The Tory vote share went up by 0.8% UK-wide (1.4% in England), while Labour's share rose by 1.5%, the Scottish collapse masking a 3.6% increase in England. In fact, if you consider the broad "right" to encompass the LibDems as well as the Tories and UKIP, the vote share of the forces of darkness has fallen from 62.2% to 57.4% since the 2010 election. I see little evidence here that we're becoming less caring or more insular.

The truth is that neither of the two main parties did well in attracting the swing-voters of electoral lore. The difference between them was that Labour failed to take enough seats directly from the Tories, while the Tories took the lion's share of the LibDem seats, reflecting the former third party's historic representation in better-off rural and suburban areas (the loss of Torbay or Twickenham is hardly a set-back for socialism). For all the pre-election talk about UKIP being a drain on both main parties, the reconfiguration on the right was the net loss of 4.4 million LibDem voters who indirectly broke 3 million for UKIP (i.e. presumably back-filling former Tory voters seduced by Nigel Farage) with the balance split evenly across the Tories and Labour (both up by roughly 0.7 million net). In the event, it appears that the Tories were the party pursuing a 35% strategy, with their success in pushing this up to 36.9% enough to deliver a working majority.

The scale of the SNP's victory in Scotland has led to blather about a "one-party state", but I suspect the higher turnout (71.1% versus 66.1% for the UK as a whole) masks a difference in behaviour rather than unanimity of belief. It looks like Labour's "get out the vote" strategy failed, and did so largely because not enough time had passed since the referendum, which in turn also maintained the momentum of the SNP. In the South, abstention is too often framed as a narcissistic consumption preference (e.g. Russell Brand), even though it is the poor and the young who generally don't vote and their behaviour is often driven by negative rather than positive reasons (insufficient bandwidth, lack of self-esteem, atomisation etc). In the urban centres of the North of England and Scotland, it is clearly a sign of working-class disengagement rather than any counter-culture affectation.

Democracies do not naturally produce one-party states, not just because parties inescapably reflect distinct class interests, but because careerism encourages competition. The question for Scotland is what will happen first: Labour revives or the SNP splits between left and right under the pressure of domestic politics. The problem for Scottish Labour is that a perceived shift to the right by the UK party will inevitably be cast as "Blairite" north of the border, regardless of the personnel or policy specifics, so inhibiting any local revival, though it's worth remembering that Blairism as socioeconomic policy was not unpopular with many Scots before the Iraq War (they're not all Scandi social democrats). The SNP will use both the prospect of independence and opposition to the Cameron government to maintain unity, but "standing up for Scotland" must eventually give way to enacting domestic policies that will divide Scottish opinion.

A prominent theme of the immediate post-election coverage has been the unreliability of the opinion polls, with the grave news that there is going to be an "independent inquiry" (a piece of laughably self-important nonsense by the British Polling Council, an industry booster). I'm not surprised that the polls were out, simply because polls exist to generate and frame debate. They are propaganda tools. The clue is the word "opinion"; we just need to remember that the opinions that drive these surveys are those of the media owners and managers who pay for the polls, not "the people". The air of polite dislike that you can sometimes sense on the TV between journalists and pollsters reflects a power struggle within the media between different professional groups competing for status. On Thursday, this spiralled into outright contempt, which was at least entertaining.

Survation gave the game away when they admitted they had "chickened-out" of publishing a poll on Wednesday that gave the Tories a 37% to 31% lead because it was "out of line". I suspect the problem with such polls is less systematic bias (the famous "shy Tory" problem) than the gulf between opinion and intention. Most people are polite and don't like to disappoint, so if pressed we'll give an opinion, even if it's a white lie. But that doesn't mean we're going to act on it. I think Leighton Vaughan Williams (impressive moniker) is close to the truth when he says "In this case, it seems a very reasonable hypothesis that rather more of those who declared they were voting Labour failed to actually turn up at the polling station than was the case with declared Conservatives". Bloody shirkers.

To give him his due, the battily earnest Allister Heath did sense something was up earlier in the week when he noted that City betting had shifted towards the Tories. Naturally, he went all Hayekian about it: "Individuals with real, exclusive local information are disproportionately likely to place bets, which means that the ensuing prices tend to be remarkably accurate; like all other markets, political spreadbets are better at creating and marshalling bottom-up knowledge than even the best top-down pollster." This is the classic dichotomy of markets, as interpreters of dispersed knowledge, versus the planned economy of centralised polling. The obvious question is: what is the "bottom-up knowledge" that those placing these bets have privileged access to? It's not like getting a racing tip from a stable-lad, unless you believe that City types have a direct feed into the inner thoughts of the voters of Nuneaton (care of GCHQ, presumably).

The most plausible explanation is that City workers are simply following the wider pattern of betting, because that's what many do for a living. But while professional investors (and bettors) know that what matters is the aggregate opinion of the market, amateurs place excessive weight on their own opinion, believing that they can buck the market when the real money-making trick is to exploit the market's movement (as Rupert Murdoch has shown). Betting on political parties then becomes an expression of confidence, rather than a sober assessment of the likely outcome, however that expression of sentiment can indicate differences in intention. Unlike betting on a horse-race, a bet on a general election result will be influenced by the individual bettor's own action - i.e. the intention to cast a vote. Though small in isolation, an aggregate shift in sentiment can therefore point to a large enough shift in intention to materially affect the outcome.

The other tiresome theme of post-election coverage has been disproportional representation: the difference in the number of votes each party needed to secure a seat (i.e. average votes per seat). This is a statistical artefact that tells us nothing about "fairness", but inevitably gets wheeled out at such times in order to advance the cause of electoral reform (i.e. the cause of centrist stability) and to console the smaller parties. That 3.8 million UKIP voters only produced a single seat is a distraction from the desertion of 4.4 million LibDem voters that produced a Tory majority. The bottom line is that Labour didn't turn up in sufficient numbers in key marginals, while too many LibDems did turn up but voted Conservative and thus negated the loss of Tory supporters to UKIP. While Farage & co will have eroded Labour support in some seats, they were probably no bigger a problem than votes lost to the Greens, and both of these trends would have been swamped had the LibDems switched in greater numbers to Labour. The story of the election is that the LibDems put David Cameron into Number 10. Again.


  1. Picking up on your point about the polls being wrong.

    A lot of people are seriously put out that the result did not conform to prediction (separate from grief over the Tories winning). Not least the media who seem to have set up studios in Westminster to cover coalition negotiations which will not now take place. Polly Toynbee has been led astray by polls that were "outrageously wrong".

    Looking at this

    It's entirely possible a lot of people only made up their minds close to the poll. Looking at the complexity our political setup the expectation that the polls should be accurate would appear to be misplaced just from technical considerations. I can see you could get the National vote share of the big parties accurate to a few percent either way from polls, but getting the seat count right with the collapse of the Lib Dems in a first past the post system, in hindsight at least, looks unlikely. The exit poll with a large sample size and questioning people outside the polls is different. I have no inside knowledge but most of these models are probably only spreadsheets (ok you can do a lot with spreadsheets). It's not like the weather where a Met Office super computer makes a difference. There has been a lot of money spent on ringing people up and asking questions. They probably haven't spent much on the actual models that they use to predict seats.

    People had unrealistic expectations that the result could be predicted accurately. Is this an example of a more general phenomena. Unrealistic expections of technology in general perhaps.

    One of Lord Ashcrofts aims in doing so much polling was to show the Tory party what a bad job Lynton Crosby was doing. It was never a surprise for him to issue polls saying labour had a chance in the marginals. I guess he was hoping to get more influence in the campain at some point. That hasn't worked out as planned for Lord Ashcroft. Lynton Crosby's consulting fees will now be astronomic if they weren't before. All the consequence of something possibly the equivalent of a throw of a dice.

    1. "Unrealistic expectations of technology" play a part, but I think this is evidence of a wider obssession with quantification - the belief that everything can be measured and thus targeted - which is a feature of neoliberal ideology.

      You can be sure the consequence of this "debacle for the pollsters" will not be that polling goes out of fashion, but that there will be calls for even more sophisticated data analytics (and Theresa May will get her snoopers charter through Parliament).

    2. Thanks for that.

      There may be a chink of light.

      Listening to Theresa May explain what an extremist organisation was on the radio today seemed to match my idea of the Tory party quite closely. As soon as this legislation gets royal assent we could use it to lock up the entire cabinet. All we need is one progressive Judge prepared to use the law in a flexible way. I have contacted the Law Society but disappointingly they tell me Judge John Deed is a fictional character.

  2. Mr Timoney, Sir.

    Is there any unlikely chain of events that could see you stand for leader of the Labour Party. The parade of neo-NewLabourites across the TV since Blue Thursday has been dispiriting. I hope I could speak for at least a wafer thin majority of your readers when I say we could offer you qualified, sceptical support for such an endeavour.

    Obviously should Arsene Wenger step aside for some reason you would have to resign your political commitments and take your rightful place in the dugout at the Emirates. Unfortunately it would be impossible to combine the role of Arsenal Manager with Leader of the opposition/Prime Minister. Managing Arsenal is a serious full time job.

    1. If Boris Johnson can combine being Mayor of London with being an MP and cabinet-member without portfolio, then I don't see why I couldn't do both jobs. I may have give up the paper-round, mind.