Show Hide image

What I got right, and wrong, in 2019

Another year, another set of howlers.

My biggest mistake of 2019 actually happened in 2018. When Boris Johnson quit Theresa May’s Cabinet I wrote that it was “far more likely that Johnson’s exit marks the end of the former mayor’s frontline career rather than May’s”. The headline went even further, describing his resignation as “the end of his career”.

In December 2019, Boris Johnson is Conservative Prime Minister with a majority of 80. What did I get wrong? The short answer is: I listened to Tory MPs, but I didn’t think hard enough about what they were saying.

As I wrote at the time, Johnson was “deeply divisive” – this is a coded way that journalists say “widely disliked” – among Tory MPs. My conclusion, based on conversations with them, was that they had decided that his error-strewn tenure as Foreign Secretary made him unfit to be Prime Minister and that there were several organised attempts to prevent him becoming so, and therefore he would not be.

This was a pretty widely shared error – in fact, I think it’s probably the mistake I’ve made in analysing British politics where I’ve had the most company ­– but I think it was also my worst one. Think about it for a moment: the argument that Boris Johnson was finished was as follows: “Conservative MPs think that Boris Johnson is an election winner, but they think he would also be a terrible Prime Minister. Therefore, he has no chance of being Prime Minister”.

Actually think about that: no chance? There was never going to be a time when, facing bad opinion polls, a dodgy economy, a popular leader of the opposition, a split party, or whatever misfortune you care to imagine, that Conservative MPs would go, "Okay, yes, I can’t stand the guy, but we’ve got to risk it?"

If anything I think the correct prediction wasn’t that Boris Johnson had “no chance” but that he had a close to 100 per cent chance of becoming Prime Minister at some point. There was always going to be at some point a crisis facing the Tory party and they were always going to reach for the person they thought was best-equipped to get them out of the hole, which was Boris Johnson.

What can I learn from this? Well, what I’ve done since is two things.

The first is that, ultimately MPs – and indeed people in general – are terrible at predicting what they will actually prioritise when push comes to shove. If in 2016, you’d told me that ten Conservative MPs would voluntarily leave the party during the Brexit process, I wouldn’t have struggled to name ten: but I wouldn’t have included the likes of Nick Boles or David Gauke, and would have included Nicky Morgan and Philip Hammond.

Some Conservatives – including ones in ultra-marginal constituencies – opposed Johnson to the end. They thought he was a winner, but they also thought that his inability to do the job meant he had to be stopped. An MP saying “in the event of x, y, and z I will vote for something” is not particularly reliable. An MP saying: “Today, I will vote for this specific measure” usually is.

Applying these measures worked pretty well in the back half of 2019. I still correctly predicted that a majority existed for Boris Johnson’s Brexit deal, and my general scepticism about the willingness of MPs to risk their party position meant that I got almost every Brexit vote in the Commons right – but I want to keep an eye on this in 2020, because I don’t want to overcorrect in the opposite direction. There are problems with tweeting out every half-baked source quote to your followers without applying anything resembling scrutiny beforehand, but there are also problems with responding to everything MPs tell you with a heavy dose of “Yeah, and the three bears”.

I thought Boris Johnson would put a customs border in the Irish Sea

“If, as this government does, you want free ports and a deep and meaningful US-UK trade deal, the only policy which ticks all your boxes is a thicker regulatory border in the Irish Sea,” I wrote in August of this year.

Policy is actually pretty easy to predict if you take it seriously – there were only two available solutions to the question of the Irish border, and one (customs and regulatory alignment for the whole of the United Kingdom) would prevent the United Kingdom reaching for what many Conservative Brexiteers see as the benefits of that.

My underlying bias here was, as I tweeted at the time, my assumption that, fundamentally, most people aren’t idiots, and that eventually the Conservative party would pick the only available option to achieve its policy aims. In general I think this approach works well.

But I didn't think the 2017 parliament would pass it. I hugely overestimated the amount that Conservative Brexiteers cared about the union between Great Britain and Northern Ireland. I correctly wrote that it would pass once it existed and was on the table, but my assumption was that a new deal based on a moved border would have to wait for an election to pass. Instead, it passed second reading.

I think this comes back to the same lesson about Johnson's leadership hopes: people in general are rubbish at telling you what they really care about until they're forced to decide. Some of the Conservative MPs who thought they could live in a no-deal party couldn't. Some MPs who thought they'd quit the party if Johnson became leader ended up in his Cabinet. And in the end, Conservative Brexiteers decided they cared more about a high-divergence Brexit than customs checks in the Irish Sea. 

I thought that Boris Johnson would be able to survive a Brexit delay past 31 October unscathed

One very silly idea doing the rounds in the autumn was that, if the United Kingdom was still in the EU after 31 October, the Conservatives’ popularity would fall off a cliff and the general election would become more winnable for the forces of Remain. This is why the various opposition groups delayed going to the country until after Boris Johnson had negotiated a Brexit deal that could command a parliamentary majority – don’t forget, his deal passed at second reading in the last parliament – at which point their various positions, of opting to revoke or have another referendum, weren’t in contrast to a no-deal Brexit (which has always been highly unpopular) but to a Brexit deal.

At the time I wrote that, “I still essentially assume that the average voter is a lot more fairminded than political partisans would like”, and that, just as a particularly silly argument about why Jeremy Corbyn wasn’t going for an election hadn’t moved the polls against Labour, the opposition parties locking Boris Johnson in a cupboard so he couldn’t seek an extension would, likewise, fail to dent perceptions of Johnson.

I’m not saying that I think the average voter always reaches a fair conclusion, and one crucial difference in the Brexit delay row is that it is one of the – perhaps the only – parts of Brexit and the 2019 election in which the BBC covered the actual granular detail well.

I think a passing observer, who gets their news in snippets on music radio and from the BBC’s social media presence would have known full well that Johnson had sought to suspend Parliament, mislead the Queen and kick out 21 of his own MPs. I’m far from confident that a passing observer would be able to tell you about the big and radical changes that his Brexit deal will mean for our economic model, or the big and radical implications of Jeremy Corbyn’s plans for the labour market.

I thought the Conservatives would win the election

Not particularly impressive this: so did everybody else who tried to approach it from a position of fairness. No-one who contributed to our constituency reports came back expecting anything other than for Labour to lose votes and seats.

I also expected them to avoid wipeout in Scotland, which they did. This was – I think – again, a pretty universal belief that everyone who spent any meaningful amount of time in Scotland came away with.

What I find heartening about both of these predictions – and my correct belief that the Liberal Democrats would be wiped out in Wales – is that it shows that constituency reporting, properly done so you don’t just talk to pensioners and other people around in the middle of the day, can properly inform you about what goes on. So too can looking at endless spreadsheets of the leave vote in specific Scottish constituencies.

But I also thought the Liberal Democrats would do better

I thought – and wrote repeatedly – that in so decisively turning the page on David Cameron’s approach, the Conservatives were risking a major Liberal Democrat revival which could cancel out gains made in the Conservative-Labour battlegrounds. As I wrote shortly after the election of both Johnson and Swinson to lead their respective parties:

“The big Conservative hope is that, while Swinson will cost them seats directly, she will indirectly tip Labour seats into their hands in Labour-Conservative battlegrounds – if the Liberal Democrat vote goes up at Labour’s expense and the Tory vote holds steady. Is the gamble right? Either way, that the Conservatives have picked Johnson, whom many liberal voters dislike, as their standard-bearer means that they have little chance of ending the Liberal revival – they can only hope that they lose less from it than Labour. Whether they realise it or not, the politician they are placing their hopes in isn’t Johnson but Swinson.”

This election was the reverse of 2017 for the Liberal Democrats: in 2017, the party got its worst ever performance in terms of votes cast but increased their parliamentary representation by a third. In 2019, they increased their vote share by a third and ended up with fewer parliamentary seats than they got in 2017. The Conservative gamble worked.

What can I learn from this? I’m inclined to put a pin in this one to be honest. The Liberal Democrat vote didn’t have to be particularly more efficient to change the outcome of the election. If just 20,000 Liberal Democrat votes – a touch over 0.5 per cent of the total Liberal Democrat vote – had been better distributed the party would have picked up 14 more seats directly from the Conservatives and saved more than a dozen seats for the Labour party. That would be the difference between a Tory majority of 16 and the current majority of 80.

I have a number of theories about why this didn’t happen but I am intensely reluctant to draw any big conclusions about what I got wrong in analysing the Conservative-Liberal Democrat battleground given that the margins here are so small. You can draw any number of perfectly strong data-driven conclusions as to what went on here: perhaps if the Liberal Democrats hadn’t backed revoke, or if the broadcasters hadn’t allowed a Johnson vs Corbyn head-to-head debate, or the Liberal Democrats hadn’t spent time and money failing to win Kensington, Cambridge or Vauxhall from Labour, they might have held Carshalton and Wallington, and won South Cambridgeshire and Guildford from the Conservatives. But a multiplicity of things that could have been different makes it tricky to say with confidence what I think I should have done differently, analysis-wise, though I will review this after the next election – and it will, I’m sure, be something I revisit in my postmortems at the end of 2020, 2021, 2022 and 2023, particularly after the devolved elections in 2021.

The election reiterates that the Liberal Democrats do better in elections where Downing Street is not in play and people are less worried about wasting their vote by backing the Liberal Democrats, like local elections or (RIP) elections to the European Parliament. But we knew this already.

I was right about the government’s new Brexit message

Downing Street briefed pretty heavily that it would continue to use divisive terms like “Surrender Act” as much as it pleased and that the message was incredibly effective. This worked very well because no-one else seemed to pick up on the fact that the only place you could see the government using the language of surrender and division over Brexit was anonymous sources insisting that they were going to keep using the language of surrender and division.

In public, however, their party conference was a full-on reboot to a message I thought worked very well, something I was the first person to notice, or at least, I think to write about: Get Brexit Done. As I wrote at the time:

“Get Brexit Done” has multiple benefits. The only Brexit proposition that still commands majority support, across the Remain-Leave and left-right divide is that people want the Brexit process to be over. So it makes sense for the government to prosecute the argument that the only way for it to be “over” is to give the Conservatives a majority – and has the benefit of being considerably less divisive than their previous “a nudge, a wink, and a not-so-gentle-hint that the other lot are all traitors” approach. It also slightly dedramatises the 31 October leaving date, which the government is highly unlikely to meet.”

As I continued onto to say, the “Get Brexit Done” phrase was very clever: but I had one question: was the pivot too late in the day?

This troubles me, because one of the conclusions I came to in my 2017 postmortem was that election campaigns really did matter, and that therefore of course, changing your message before the campaign had actually started was fine and dandy. This bothers me because the point of my annual postmortem isn’t to have a fairly easy to put together piece over the Christmas holidays (although that is certainly a bonus) but to improve and sharpen my analysis.

The big thing I’ll change this year is to be more aware and explicit about how my previous mistakes have changed my analysis – so I can be more sure I’m actually learning from them.

Stephen Bush is political editor of the New Statesman. His daily briefing, Morning Call, provides a quick and essential guide to domestic and global politics.