Note that like any sensible person I have given up making political predictions and, to be clear, what appears below is not one. However, as we pass the twentieth anniversary of Tony Blair’s Labour Party enjoying the biggest post-War landslide election win ever (an absolute majority of 179), it is worth presenting the case for why the Conservatives will not enjoy anything like that kind of margin on 8 June.
Let us start with the dangerous assumption that the polls are roughly correct. Work has been done since they underestimated the Conservative vote share by roughly four points in 2015 to correct certain assumptions made, and it should be noted that even in a bad year they are never so well out as to suggest anything other than a comfortable Conservative lead on polling day. There has also been considerable work on polling for people switching vote (to which we will return) noting that 2016 Leave voters and particularly among those 2015 UKIP voters are indeed intending to switch to the Conservatives. Therefore, the current poll of polls suggesting that the Conservatives are currently on around 46% in Great Britain and Labour on 29% is a reasonable, if probably not absolutely accurate, starting point.
(It should be noted that the figure of Conservative around 46% and Labour around 29% is universal across all class backgrounds. The notion that the Conservative vote is predominantly middle class has long been flawed, and is particularly so post-referendum. It is in fact the Liberal vote which is clearly predominantly middle class, and has been for some time; and the UKIP vote which is predominantly working class.)
There are three prime reasons Labour may not come out quite so poorly (and the Conservatives may not come out quite so well).
The first is in the figures themselves. Unmentioned in the top-line figures is that polling is picking up an unsurprisingly but notably high number of ‘undecideds’. Of further note is that these ‘undecideds’ are particularly weighted towards people who declare that they voted Labour in 2015. This suggests that a very high number of Labour voters last time are unsure they wish to do so again in 2017 (in terms of polling, we should make no assumptions about why that is). However, there are very few direct Labour to Conservative switchers even among those declaring a preference (the Conservative uplift is primarily from UKIP and non-voting); therefore while we should not assume that these Labour-leaning undecideds will necessarily break Labour, we can be fairly sure they will not break Conservative. In other words, the Conservative figure of 46% is the highest they could possibly score (as the best case scenario for them is that current ‘undecideds’ break evenly or do not vote at all).
The second is that in every case where a ‘snap election’ has been called in post-War Britain (and generally in other comparable Western democracies), the governing party has lost a disproportionately high vote share in polling during the campaign itself. Typically the vote share at the start of the campaign is fairly accurate (at least insofar as the polls themselves are) as most people make up their mind between elections, but a ‘snap election’ gives them less time (hence it is unsurprising to find a higher than usual number of ‘undecideds’). What appears to happen is that supporters of the party calling the snap election declare their support quite contentedly from the start (perhaps on the assumption that they would not have called it if they were not reasonably sure of winning); however, people not pre-disposed to voting for the governing party take longer to make up their minds (since they have, by definition, not had a full electoral cycle to settle on a particular party preference).
The third (and all of these reasons are linked) is simply that the Conservatives, having absorbed much of the UKIP vote, cannot possibly benefit from tactical voting (except, perhaps, in Scotland). On the other hand, all the broadly ‘open/left’ parties can expect to do so – people who definitely do not like the Conservatives may not be sure whether they really prefer Labour, the Liberal Democrats or the Greens, but they may well be willing to vote for whichever one has the best chance of defeating the Conservative. In other words, even if the polling numbers to end up something like Conservative 46% Labour 29% (a considerably larger margin than Labour 43% Conservative 30% as was the case in 1997), the Conservatives would come nowhere close to the sort of landslide Tony Blair enjoyed.
In other words, the chances are that the Conservative figure of around 46% is already a little inflated; it is likely that during the campaign it will slip (whereas the campaign rarely makes much difference); and in any case tactical voting may see them win fewer seats than they may typically expect for whatever vote share they achieve.
To be clear, I predict nothing but even the above suggests nothing other than that the Conservatives will be comfortably the largest party on 8 June. However, the current betting suggests that the even chance is a Conservative landslide majority of around 124. I am not only predicting nothing but I was never a betting man – if I were, however, I’d be inclined (just inclined, mind) to go a little lower than that.