Thread by @robfordmancs

Old Bexley and Sidcup by-election today. Labour would need a 21 point swing to take the seat. Some historical context:

Lab Con by-elec loss - Chesham & Amersham - swing 25.2 pts

Last Lab by-elec gain from Con - Corby (2012) - swing 12.7 pts

Other swings in post Brexit Con defences:
Brecon & Radnorshire (2019) - 12 pts (lost)
Sleaford & N Hykeham (2016): -0.3pts (won)
Richmond Park (2016): 21.7 pts (lost)
Witney (2016): 19.3pts (won)

Average swing against Con across all post-Brexit by-election defences: 15.6 pts

There were also huge swings against the Cons in 3 2014 by-elecs, the last defences before Brexit:
Rochester & Strood - 28.3 pts (lost)
Clacton - 44.1 pts (lost)
Newark - 15.5 pts (won)

So while a 21 point swing seems on paper a huge ask, larger swings against the Cons have in fact happened in 4 of their 8 most recent by-election defences, and two others came pretty close.

I do wonder if perceptions of Con strength in by-elections end up being skewed by the relatively much higher frequency of Labour than Con defences in recent decades.

This is only the 10th Con seat being defended in a by-election since the current period of Con govt began in 2010. In the same period, Labour has had 25 seats to defend in by-elections.

By contrast, there were 22 by-elections in Con seats in the equivalent period of 1979-1990, and Cons lost 10 of them.

In the decade from 1987-1997, there were 18 Con by-election defences, and Cons lost 15 of them, including every one held from May 1989 onwards