Can we talk about UK polling for a second?
The Tories may well get a bounce out of Boris Johnson's Premiership, but history cautions against over-selling it.
How to convert a snap election to a Parliamentary majority
The Takeaway: It’s common for newly-elected prime ministers to cause a polling bounce for their party. We’re very likely seeing one with Boris Johnson, but the Tories look worse under his early leadership than during Theresa May’s bounce. Ms May went on to a minority government in a June 2017 surprise election—could BoJo be in for a similar fate?
Editor’s note:
Thanks for reading my thoughts on this subject. And thanks for subscribing! Your membership adds up and makes all this newslettering possible (reminder: I do all this work independently). Please consider sharing online or with a friend; the more readers, the merrier. Remember that, apart from getting special articles, subscribers can also comment below each post and participate in exclusive threads.
As always, send me your tips about what you’d like to read about next. Or what you don’t want to read. Or your feedback otherwise. Also, cat pictures are nice, so please send them to me! I’m elliott@thecrosstab.com, or @gelliottmorris on Twitter.
Thanks all,
—Elliott
What’s happening in Britain right now
If you are unfamiliar with the state of British politics today, allow me to offer a one-paragraph refresher. The previous prime minister, Theresa May, resigned in June amid an outcry over her (failed) policy proposals to resolve the country’s long-awaited exit from the European Union (Brexit). With Ms May gone but the Conservatives still in power, they were allowed to elect a new leader for their party who would automatically become prime minister. While the party was in turmoil, their support in public opinion polls began to crater. The Brexit Party had become the most supported in the country (see my chart for The Economist below) and the two parties had a historically low combined vote share.
Why not tweet out single poll results and enjoy the retweets?
With Boris Johnson now in power, the Tories are experiencing a polling bounce. This is likely because those who were supporting the Brexit Party found in the new Conservative leader a pro-Brexit prime minister. It now appears that the party has rebounded from their low in June.
But it is uncertain what will happen next. Will the BoJo Bounce turn into the BoJo Bump?
Listen, I like tweeting out single polling results and extrapolating future elections enough as the next guy, but that seems like an exceedingly bad idea in Britain right now.
In the first place, the two major parties are essentially tied. Although YouGov released a poll yesterday that showed BoJo up 10 points on Labour, other polls disagree and put Tory vote intention at as low as +2 (within the margin of error for a tie).
In the second place, it is wholly unclear how predictive even the average of polls today would be. When Ms May first took power in 2017, the Conservatives leapt to a possibly 16 point lead over Labour. Emboldened by the results, she called a snap election—and went on to lose a majority of seats in Parliament.
I guess I’m climbing up on my hobby horse now in saying that the future is uncertain and you shouldn’t pay attention to single polls, but…. the future is uncertain and you shouldn’t pay attention to single polls! Even if we formalized a model about election outcomes, it would be showing a ¯\_(ツ)_/¯ emoticon (possibly with the Conservatives slightly ahead of everyone else). We have seen how close vote intention and temporal concerns make predicting a contest tough.
As Rob Ford pointed out in The Guardian yesterday, a third source of uncertainty is the geographic distribution of Brexit Party, Conservative, Labour and LibDem voters. It could be that the Tories lose a lot of seats in a snap contest because Brexiteers are concentrated in rural and suburban constituencies and can win many seats there due to first-past-the-post plurality voting. Or it could be that the Liberal Democrats wipe out Labour in London and the Union strongholds in an accentuation of what happened in this year’s elections to the European Parliament.
Looking to Martin Baxter’s Electoral Calculus for advice, I can provide a reasonable range of outcomes:
IF election-day vote intention matches today’s Britain Elects average:
Con: 25% of the vote, 225 seats
Lab: 25% of the vote, 252 seats
LibDem: 19% of the vote, 65 seats
Brex: 19% of the vote, 47 seats
IF Tories expand their margin to today’s most optimistic scenario (+10)
Con: 31% of the vote, 358 seats
Lab: 21% of the vote, 183 seats
LibDem: 20% of the vote, 51 seats
Brex: 13% of the vote, 0 seats
IF Tories collapse (like pre-BoJo polls showed, -3)
Con: 23% of the vote, 181 seats
Lab: 29% of the vote, 304 seats
LibDem: 17% of the vote, 56 seats
Brex: 19% of the vote, 48 seats
If you want a simpler look at how Labour’s vote swing relates to their swing in seats, see this graph from Martin Baxter:
Since the polling average is so close and possibly not at all predictive, it seems to me that we’re experiencing the forecasting equivalent of ¯\_(ツ)_/¯, and so I don’t think any tweets or articles predicting the contest either way are reliable whatsoever. This is especially true when we consider that the next contest could be 5 or 15 months away and we don’t know what the political and economic factors influencing vote intention will be.
As I said on Twitter, we should remember the mistakes of the past—and the pain suffered by any person who predicted a huge Tory majority the last time they got a new PM and a bounce in the polls:
