Politics: punditry

By Richard North - February 17, 2024

To say that Thursday evening and the early hours of Friday morning constituted “another disappointing election night for the Conservatives” – as does John Curtice – is rather like saying that breakfast service on the Titanic on 15 April 1912 was not up to standard.

This is where trying to analyse and understand by-elections gets so difficult. It is not the events themselves so much as the punditry that comes afterwards that muddies the water and can leave one confused and uncertain.

But when it comes to statements of the bleedin’ obvious, that is one of the joys of being the BBC’s favoured pundit. You can come up with any old tosh and the presenters will nod wisely and commend you as the great expert to their dwindling audiences.

The other waste of time comes with the insistence on prattling about “swings” when analysing the results. Fairly obviously, though, when turnout is down – as it so often is in by-elections, even relatively small changes in voter sentiment may yield exaggerated results.

Thus, when Curtice prattles about a 28.5 percent swing to Labour at Wellingborough – the second highest swing from Conservative to Labour in any post-war by-election – he completely misses the point that the by-election candidate, Gen Kitchen, attracted only 13,844 votes compared to 13,737 votes in the 2019 general election – an addition of a mere 107 votes.

Since comparative turnouts between elections are never expressed by party – and never can be since we simply don’t know how individuals voted – there is actually no way of telling how many more people actually voted for Labour than they did in 2019 – and how many of the 2019 Labour voters dropped out or changed their votes last Thursday.

In other words, the “swing” figures are completely meaningless and there is absolutely no point in using them. They tell us nothing.

Hence, when we see Starmer downplaying the results, saying that Labour has “more work to do”, he is not only being sensible in trying to guard against complacency, but is also probably aware that the Wellingborough result is not necessarily a great harbinger of a general election victory.

Looking at that result in the round, even with a number of uncertainties, it is entirely reasonable to interpret the massive drop in the Tory vote, from 32,277 in the general election to the dismal 7,408 votes cast last Thursday, as a “stay-at-home” vote.

In fact, never before in the history of the constituency has the Tory vote dropped that low. Even in the landslide 1945 Labour victory, the Conservative candidate took 16,426 votes and since 1970 until this by-election, the Tory vote has not dropped below 20,000.

Thus, what we are undoubtedly seeing is the reverse of the 2019 election where we saw Labour voters deserting in their droves. Rather than Johnson winning the contest, Corbyn’s Labour lost it.

This dynamic should profoundly worry Sunak. Despite many pundits actually agreeing that the voters showed little enthusiasm for Labour, he failed to energise his voters and they stayed at home rather than go out and vote for him.

That said, there are aspects of the Wellingborough contest which render it untypical, even for a by-election. As the Guardian points out. The election was triggered by the recall of disgraced MP Peter Bone.

But, in what seems from the outside to be an inexplicable decision, the constituency chose Bone’s partner, Helen Harrison, as his replacement. The Guardian cites a “Tory activist” who said: “None of us wanted to campaign in Wellingborough, we just couldn’t face having to justify our choice of candidate”.

If that was the case, then there was a personal element to this by-election which is unlikely to be replicated in the general election, which substantially reduces the predictive value of the result. Thus, the paper thinks that Kingswood could be more representative of national opinion.

On that basis, it suggests the Labour lead is softening slightly after three byelections last year where the swing was more than 20 points on each occasion. In this case, where swings are being compared between recent by-elections rather than the long-term historical trend, they may have more value, with Labour’s Damien Egan delivering only an 11.5 percent swing.

Here, we are told, the Tories are trying to put a gloss on Thursday’s defeats by highlighting the low turnout in both seats, pointing to the 37.1 percent in Kingswood and 38 percent in Wellingborough.

These are said to be below the mid-40s seen in recent byelections in Mid Bedfordshire, Somerton and Frome, and Selby and Ainsty. But then Tamworth in October 2023 – at the same time as mid-Beds – only delivered a turnout of 35.9 percent in a gain for Labour against the Conservatives.

Taking the last 21 by-elections, from 2021 through to the very latest contest, we have seen turnouts range from a high of 52.3 percent (Tiverton and Honiton in 2022) to a low of 25.8 percent (Stretford and Urmston 2022), with the average standing at 39.2 percent. Thursday’s by-elections were slightly below average but well within the normal range.

What was significant about Kingswood, though, was that while the turnout was roughly half that of the general election (37.1 as against 71.5 percent), the Tory vote dropped to just under a third. The Labour vote, however, only dropped by a third to bring Egan his victory. Here, therefore, are indications of a Tory stay-at-home vote, albeit not as extreme as that experienced in Wellingborough,

Thus, if Kingswood is more typical of what might happen in the general election, then the Tories really are in trouble. This is the constituency where there was a discernible “Ukip effect” so it is arguable that a combination of a switch to Reform UK and the stay-at-home vote cost the Tories this seat, a dynamic that could be repeated at the general.

There is one further local factor – the collapse of the Lib-Dem vote. Normally, that plays to the advantage of the Tories – as it did in the 2010 general election and the 1951 election. Coincidentally, in that earlier election, the disappearance of the Liberal candidate, whose vote had deprived the Tories of the seat in 1950, gave the seat to the Tories in Wellingborough.

It would appear, therefore, that the combination of circumstances in these two by-elections worked against the Tories, to the advantage of Labour. But, at a general election elections, those effects – if still present – will not be replicated uniformly on a national scale.

For instance, as in 2015 – where the collapse of the Lib-Dem vote cancelled out the “Ukip effect” – in Tory marginals which have previously had a strong Lib-Dem presence may be rescued by the absence of a third party challenge, especially as we have probably seen “peak Reform”.

Then, on a national scale, there is Labour’s Muslim problem – which had no discernible impact in these by-elections. Depending on how the Gaza situation pans out by the time we have the general election, a Muslim boycott could cost Labour enough seats to deprive it of victory – especially if the SNP vote even partially holds up.

To an extent, therefore, these by-elections may give no guide at all as to the future which could just as well be dominated by “events, dear boy”. For instance, if the November ONS figures on immigration don’t show a substantial drop on last year – which may well be the case – that could destroy the Tory’s chances of re-election.

Just as easily, if Sunak’s Rwanda policy self-destructs – which also seems likely – that could be curtains for the Tories, given that immigration holds up as the main concern of the electorate.

As far as the chatterati in The Times go, though, the general election is a done deal. Starmer’s difficulties are “all noise”. Labour will win the election.

Nevertheless, there is that old English saying: there’s many a slip twixt cup and lip – as Starmer is finding out. Others might say it ain’t over until the fat lady sings, and there’s no fat lady in sight just yet.