Well, I started well. “My strong expectation from travelling the country and talking to campaigners is that Jeremy Corbyn will beat Ed Miliband’s vote share in 2015,” I wrote. True. But then it all started to go downhill. “[Corbyn] may even match Tony Blair’s in 2005”, I declared – he went on to exceed it by five points.
But I wasn’t done. “But I also think that these extra voters are insufficiently distributed thanks to first past the post, and that the party will lose significant numbers of seats.” Labour gained 33 seats and lost just three.
I’m going to write more about what I missed and why in my end-of-year review, as I did in 2016 and 2015. However, as there is the possibility – not at all likely but plausible in my view – that there will be another election before the year is out, here are some thoughts on what I missed that will, I hope, inform how I cover elections to come, particularly if the next one comes before its expected date.
Campaigns do matter – sort of
“Theresa May will not have a good campaign by any objective standard,” I said at the start of the election campaign. “Her interviews will be halting, if she does come into contact with any actual voters, it will go badly – but it won’t matter, because people fit the facts to fit their opinions of the leaders, rather than the other way around.”
Three out of four ain’t bad. Theresa May did not have a good campaign, her interviews were halting, her interactions with real people tended to end badly. But it mattered a great deal.
(The useful thing about the local elections is to give us a pretty good indicator that Corbyn’s good campaign and May’s bad one really matter. Don’t forget that at the start of the contest, the Conservatives had romped to victory in the local elections. No opposition party has ever done better at a general election than the preceding local elections, so we can really see the impact of the campaigns here.)
It’s worth noting though that while this election showed the damage that bad campaigns can do, they also showed how much you can get away with before people notice. Before the social care manifesto, May had had one awkward encounter with a woman with learning difficulties, the Conservative campaign had got its sums wrong over the minimum wage and school meals, and an event intended to showcase Labour’s economic weaknesses instead highlighted divisions between May and her Chancellor, Philip Hammond.
Yet voters consistently said that May was running the better campaign than Corbyn, as they liked her more, and generally by margins of 20 points or more. It was only after the social care fiasco that the gap closed.
What I realise is that what usually happens is not that campaigns don’t matter, but that campaigns tend to cancel each other out. May’s bad campaign wouldn’t have mattered, I think, if Corbyn hadn’t been running such a good one. People had to first be turned off by the Conservatives to give Labour a look, but they also had to like what they saw. If the Tory campaign had been better, or Labour’s worse, it wouldn’t have mattered.
The Conservatives also had to consistently fail to match Labour’s moves. Labour ran a good campaign focused on the cuts to school spending – helped by the fact that the National Union of Teachers was also campaigning against the government’s planned cuts – and the Tories did try to head that off by fixing the problem in their manifesto. But they fixed it by diverting funds from universal free school meals towards other areas of school spending – aggravating rather than fixing the issue. The contrast with David Cameron, who chucked an 11th-hour pledge on childcare into the 2015 manifesto to spike Labour’s guns, is marked.
In future, I think I’ll get a better idea of how campaigns are going if I look at them not as isolated things but as matching one another. What was the Tory attack on Labour? Pretty muted and focused on Corbyn’s historical links. What was their promise to voters? Thin gruel.
I should have appreciated what a big deal Theresa May’s manifesto U-turn was
No major party ever in British history has abandoned a central plank of its manifesto during an election campaign – in fact, I can’t find an example of a party abandoning anything of substance from its manifesto during the election campaign.
At the point that happened, the correct response from anyone trying to analyse the election should have been, “Something crazy has happened, therefore all bets are off”.
While I expect that the next Conservative campaign will not be as maladroit as the one just gone – it could hardly be worse – I should have taken more notice of the fact it was a shock event.
As YouGov’s election model, which called the result correctly, put the Conservatives on course for a big majority before the Tory manifesto was unveiled, it’s clear that this was the decisive event in the contest. I will be more alive to the impact of shock events in the future.
I gave too much weight to private information
What I ought to have written before the election is this: “The polls are telling us two different things. Off-the-record, I am privately being told that the worst polls look closer to the truth. So you pays your money and makes your choice.”
But instead I assumed that private information was superior to public information and leant more heavily on that. In general, one thing I think journalists as a trade get wrong is to be overly excited about things with the word “private” or “secret” in front, and less excited about things that are freely available, but neglected.
We saw how Labour effectively exploited that with their “leaked manifesto” – a manifesto leak that, I’m reliably informed, came from the shadow Treasury team – which, because it was leaked, was covered in more detail – and more policy-focused detail – than almost any manifesto in British history.
Private information can be useful of course – we shouldn’t forget that the parties pay for, and collect with volunteers, a lot more data than anyone else, meaning they ought to know a thing or two – but it should always be treated with scepticism, and I’m going to do more of that from now on.
However, what I failed to appreciate was how the snap election changed things
It’s possible to take that last part too far. If Jeremy Corbyn’s aides and allies say he is gaining votes but losing seats, and his opponents, internal and external, say he is gaining votes and losing seats, you can’t really conclude anything else other than the most likely outcome is for Labour to gain votes and lose seats.
Nor was there a professional/volunteer divide. I talked to in excess of 100 activists from all parties as well as the salaried staff, and although party activists were a bit more accurate – around a third of the activists I spoke to were correct in the result on the ground, significantly better than the paid operatives at the top of any party – they were still, on the whole, pointing towards a Labour surge that was very real but geographically concentrated in a way that made it ineffective under first past the post.
In no other political contest that I have covered have the competing sides both been so wrong about the state of the race. Vote Leave and Britain Stronger in Europe both had a very clear idea of what the battleground in the referendum contest was. In 2015, the Conservatives had a good picture across the United Kingdom, the Liberal Democrats knew they were in trouble, and although Labour’s high command was badly wrong about where the battlefield was, its army of field organisers and most of its volunteers knew that something was going wrong on the doorstep.
So what to do? Well, I think this is where the fact that it was an unexpected election should have been something I thought about more.
I spent an awful lot of time writing about how both sides, but the Conservatives in particular, hadn’t expected the election and were having to set up their operations very hurriedly. (Labour had at least put themselves on an election footing, allowing them to have well-vetted candidates in place very quickly.)
But what I failed to appreciate was that this rush wasn’t just going to have negative consequences as far as their press operations, manifesto, events team and the like were concerned: but also at the business end of campaigning, it would mean that the information they had would be significantly worse than it would have been at the end of a five-year parliament.
(This is a particularly inexcusable oversight as I lost sight of the times that field organisers, when asked how their data looked, sucked their teeth in and went “Well, I’d have preferred to go into this with five years of data”.)
This last issue is particularly important if there is another election before the allotted time. It’s not just that the public-facing bits of the campaign will be missing things. Every part of the campaign will be scrambling to keep up.