Mid Bedfordshire and Tamworth by-election preview

Voters in Mid Bedfordshire and Tamworth will go to the polls on Thursday in one of the most exciting and unpredictable by-election days British politics has seen in years.

Mid Bedfordshire

Most of the media attention and market volume has gone towards Mid Bedfordshire, which looks like a genuine three-way battle between the major parties. I wrote a preview of the by-election at the beginning of September, a time when the Liberal Democrats were last traded at odds-on (1.96) and Labour (4) just edged out the Conservatives (4.1) for second place. I speculated that the odds should be more like 2.5 (40%) for the Conservatives, 3.5 (29%) for Labour and 4.5 (22%) for the Lib Dems, and the market has since moved to that configuration, with the Tories now odds-on favourites (54%), Labour in second (29%) and the Lib Dems in third (17%).

While the Conservatives represented the early value in this market, they have shortened too much for my liking and the value now lies with Labour, who at time of writing are available at 3.65 in places.

Some ‘leaked memos’ emerged from the Conservative camp and their very real ‘volunteer market research team’ on Tuesday, but these can be safely ignored. The maths involved just doesn’t add up, and the document is clearly an exercise in expectation management rather than an informative campaign update. Conservative HQ could have tried to make the memos believable, but it seems they wanted to have their cake and eat it too: to manage expectations of their own vote share while also claiming that very few of their voters were actually switching to Labour, leaving them with an overbaked product.

What these memos do get right is that the Tory vote share is likely to collapse. Labour and the Liberal Democrats ought to be buoyed by the fact that their combined vote in Conservative-held by-elections has become very efficient in recent years. When there is a clear challenger to the Conservatives, voters have become increasingly smart in distributing their votes tactically to the right party. On a percentage scale, where 0% means the combined vote was equally split and 100% means one party took the lot, the Lab-Lib efficiency score has increased significantly over time. During the Conservatives’ time in government between 1979 and 1997, the figure steadily climbed from 27% in term one to 60% in the 1992–97 period. From 2010 to 2019, it averaged 68%, and in this parliamentary term has reached a very impressive 87%. This high level of efficiency has made it possible for both opposition parties to win previously out of reach safe Conservative seats.

The issue in Mid Bedfordshire, however, is that both parties claim to be the main challenger to the Conservatives, and neither side has backed down from the fight in the last two months, meaning the anti-Tory vote share will be somewhat inefficiently split.

My thinking on the seat has therefore not really changed much since I said the Conservatives ‘could plummet to the 30–35% range and still hold on’. Their best case scenario is one in which Labour’s vote share ends up somewhere around 20–30% — not enough to win, but splitting the anti-incumbent vote enough to allow the Conservatives to come through the middle.

My by-election model, which predicts vote shares and win chances based on previous results and current polling, lines up pretty well with the odds, giving the Conservatives a 44% chance of victory, Labour 33% and the Lib Dems 22%. This further contributes to my assessment that the Conservatives should be the favourites here, but not an odds-on shot. If you managed to get on the Conservatives at 4.1 back in September, my advice would be to trade out now.

Tamworth

For Tamworth to be overturned, it would require at least a 21-point swing from the Conservatives to Labour, which has only occurred once this parliament in Selby and Ainsty, which swung by 24 points in July. In the Uxbridge and South Ruislip by-election, which was held on the same day, the swing to Labour was just seven points.

Given the sheer size of the task in overturning a 43-point majority, the market looks far too confident on Labour at the moment, with Keir Starmer’s party currently rated a 77% chance. There has not been any constituency polling in Tamworth to justify this price, and it seems like the market is operating on vibes alone.

My by-election model currently points towards the Conservatives holding on, with a 16-point swing to Labour producing a heavily reduced lead under 10 points. I am a bit wary about the model’s output here given its slight underestimation of the swing in Selby and Ainsty.

Regardless, Labour’s poll lead has fallen by about 2–3 points since July, and even if we were generous and took the Selby and Ainsty swing as our central estimate here, we would still be looking at a Labour lead of just five points. Given the lack of data from the seat, the current odds of 4.2 on a Conservative victory look very generous. In contrast to Mid Bedfordshire, I think the Tories are the value bet here, and I’m going to stick my neck out and say I’d make them favourites to hold the seat.

Best Value Bets

Labour to win Mid Bedfordshire @ 3.35
Conservatives to hold Tamworth @ 4.2