UK 2019: Win For Polls And Tories, A Shocker For The Left And Centre

After spending yesterday commenting on the UK election on Twitter, I think it's time to put some longer-form comments down about the result in light of the sorts of themes that get covered here.

The result was of course a resounding win for Boris Johnson's Conservatives, with 365 of the 650 seats.  The major seat movements were in England, where the Tories took nearly 50 seats from Labour.  There were also movements in Scotland (where the Scottish Nationalist Party took seats from both major parties), Wales (where Labour lost more seats to the Tories) and Northern Ireland (where the Democratic Unionist Party lost two seats to nationalist parties).  The Tory vote didn't increase greatly, but the Labour vote collapsed.  

There will be plenty of detailed accounts available of the characters of areas that broke to one party or another.  As with Australian analysis, when reading these always beware of the "ecological fallacy" - an electorate with a lot of voters of type X swinging to party Y does not always mean that voters of type X themselves swung to party Y.  The obvious hook (and one not so subject to this problem as some) is Brexit, and unsurprisingly strong Leave areas swung to Conservatives and strongly against Labour, while strong Remain areas swung more modestly against Labour and weakly against the Conservatives.  However the relationship between Brexit position and swing was messier in Labour's case.  Correlation-hunting has unearthed this remarkably strong link between swing and blue collar jobs, so it will be interesting to see where that debate goes.  

Simple analysis will try to pin the blame for Labour's poor result solely on either its leader or Brexit, but in my view there were too many things wrong with Labour's campaign to isolate any one overwhelming cause.  I think the leading factors were, in some order, Jeremy Corbyn's extreme unpopularity,  the party's confused Brexit message, the party's radical-left policy platform, and the party's poor handling of the antisemitism issue. Damage control attempts that seek to blame the loss wholly on Brexit should not be put up with.  Also (in an echo of Australia) a cluttered campaign has been cited, as has (and see link above) a failure to speak to the actual concerns of the party's formerly strong areas in a way that voters in those areas liked.  Much will be said about to what extent this is a general crisis for the left given the demise of its natural organising bases, and to what extent it is down to the many special factors.  (I don't think the labo(u)r project is quite so stuffed as that, but if you have to rely on out-of-the-box leaders and on the public responding to failures and upheavals on the right in an age where voters increasingly ignore or welcome them, you might not be winning very often.)  

The result also shows, again as in Australia, how indifferent and even hostile voters are to parliamentary games.  The Tories exploited this with an excellent TV ad about how the hung parliament was just "arguing about arguing" (correctly betting that the electorate would either not remember or care that the hung parliament was the Tories' fault in the first place).  Voters in the UK, Australia and the USA don't want this stuff anymore; they want to elect leaders who will get on with the job.  Johnson's many parliamentary and court defeats only improved his electability.

I have concerns about the implications for Australian Labor, which may try to read too much applicable here into a result made extreme by extreme circumstances.  The UK Labour campaign was like the Shorten 2019 campaign on steroids and the magnitude of the drubbing may give succour to those who believe the path back to government is to hide in the corner with few if any policies and are already pushing the post-defeat debate in that direction.  Much the same tactic failed in 2001 and was on track to probably fail even before the Tampa and 9/11 incidents cemented John Howard's third victory.   The Corbyn defeat is such an extreme example that it would be unwise to regard it as telling us that much we don't already know from our own election.  Save, perhaps, that the idea voters really want expansive leftism (which was given some credence by Corbyn's performance against a poor Tory campaign in 2017) has been demolished, so Labor won't win here by trying to out-green the Greens. 

The Centre And First Past The Post

The UK election was not only a failure for Labour but an even greater failure relative to opportunity for the Liberal Democrats.  The Liberal Democrats were gunning for this election off polls suggesting they would greatly increase their seat count but in the end despite a 4% swing to them, they went backwards by one on their (meagre) 2017 seat count and lost their leader.  This was despite having picked up four Conservative and three Labour defectors during the parliamentary term, none of whom were re-elected.  Another centrist project, Change UK, also flopped miserably, and overall of the ten Labour and eight Conservative MPs elected in 2017 who recontested for a different party (often as Liberal Democrats) or as independents, not a single one was re-elected.  The bright spot for third parties was the strong performance of the SNP in Scotland, but this was none too surprising given Scotland's desire to give a firm up-yours to Brexit.  

It's all very well to say the first past the post system squeezes the Liberal Democrats, which it does, and which is all the more reason why their acceptance of a doomed Alternative Vote referendum in the 2010-5 coalition was insufficient to protect their long-term interests.  But under the same system in 1983, a precursor alliance managed a quarter of the popular vote, so either there was not much appetite for centrism or the Liberal Democrats were seen as offering a particularly feeble version of it.  Centre parties and candidates should have been able to make the case that both the Tories and Labour were so unpalatable that their voters shouldn't care which of them won, yet either there were not enough really centrist voters or they were turned off in other ways.  The remarkable age divides in UK voting might have something to do with the former.   

Examples of first past the post both hurting Labour and saving them were seen.  A striking example of the former was Kensington where the Tories squeaked to victory in the heavily pro-Remain scene of the Grenfell Tower disaster with 38.3% of the vote thanks to the split between Labour and the LDP.  On the other hand, Labour was rescued by Conservative/Brexit vote splitting in some seats where it suffered massive swings, most notably Barnsley Central which the Brexit Party would have won on Tory preferences under compulsory preferential and perhaps even under Alternative Vote (the UK name for optional preferential voting).  I haven't looked in enough detail to say how much the major parties suffered or gained from first past the post in an election so lopsided that the Tories would have won it under any single-seat system.  But with 31% of Labour voters voting tactically, it's clear third parties would have done better without the "wasted vote" problem.  There were several seats where those wanting to vote tactically had no way of knowing reliably who to vote for.

Polling

Unfortunately at least two Australian journalists (Stephen Drill and on Twitter Mark Riley) have already published claims that the polls got this election wrong.  This is, quite simply, innumerate rubbish, and it is inexcusable given that there are so many informed observers who if asked would have immediately given a more accurate assessment for free.  In the case of Drill's "analysis", Drill himself had earlier claimed the election was "on a knife edge".  His article supposedly supporting this claim cited two polls, one of which showed a decisive Tory win with a twelve-point lead, and a Savanta ComRes of which he wrote:

"Another poll put Mr Johnson’s lead at just six points, with the Conservatives on 41 per cent to Labour’s 36 per cent in the Savanta ComRes survey."

I'm fairly sure that 41 minus 36 is actually five. Interestingly the Daily Telegraph report of the ComRes poll had also used the expression "knife edge" for a report based on its poll alone, though even it predicted a narrow Tory majority.  And anyone familiar with the range of final UK polls would have known the ComRes was an outlier, with most of the late polls placing the lead between 8-13 points.  Moreover, ComRes had persistently had lower average leads than most others.

Riley's complaint was that "All the major UK papers reported polls tightening with a hung parliament possible." But considering the range of polls, a hung parliament was only possible because of the range of systematic polling errors seen in previous years.  If one assumed there was not any significant average error across the 2019 polls (an assumption no-one could safely make, though it turned out to be pretty much correct) then there was no chance of anything but a clear Tory majority.

The fact is that after three consecutive failures in 2015, Brexit and 2017, this was a really good election for most of the UK polls.  Apart from a couple of outliers, nearly all the final polls were close to the mark with some, especially Opinium and also Ipsos MORI and Kantar, remarkably so.  (Opinium seems to have nailed every party after rounding, except for having the Greens wrong by 0.8).  Polls specific to Northern Ireland also did well, which was notable given that the Northern Ireland result was quite significant and polls correctly forecasting it were met with some scepticism.  The polls taken collectively did slightly underestimate the scale of the Tories' win, but polling-based seat models still forecast clear Tory wins (generally 10-30 seats short on the median estimate).  Given the complications of regional swing, tactical voting and multi-party contests, if the polls are getting results right within about 20 seats out of 650, that's a good result not a bad one.  

The narrative of recent poll failures is a compelling one but this election is a counterexample to it and should be reported as such.