Two seats changed hands on Thursday. One in Fylde, on the Lancashire coast, where the Conservatives took a ward from an independent. One in Malvern Hills, in the West Midlands, where the Liberal Democrats did the same. The kind of result that, if you blinked, you missed entirely.
And yet ConservativeHome reported the pair of gains in its rolling council by-election digest, the way it always does. Two independents out, two party flags up. The interesting question is not what these results mean on their own. It is what they mean stacked against the other forty or fifty council by-elections that will run over the coming months, and whether anyone trading on the next general election is paying the slightest attention.
The signal hidden inside a 400-vote contest
Local by-elections are not predictive in the way national polls try to be. Turnout is often under 30%, the candidate matters disproportionately, and independents win and lose seats for reasons that have more to do with a parish row than with what Westminster did that week. A single Conservative gain in Fylde tells you almost nothing about whether the party is recovering nationally. A single Lib Dem gain in Malvern Hills tells you almost nothing about whether the yellow rosette wave from 2024 still has momentum.
The signal lives in the aggregate. Election analysts have been running net-change tallies on council by-elections for decades, and the patterns, when smoothed across a year, do roughly track the national mood. The Lib Dems winning consistently in commuter-belt wards in 2022 and 2023 was a real precursor to what happened in 2024. The trouble is that the aggregate moves slowly, the sample is noisy, and the contests where parties pick up seats from independents (as in both of Thursday's results) muddy the picture further. An independent losing a seat is not quite the same data point as Labour losing a seat to the Greens.
This is, broadly, why prediction markets and polls tend to ignore the council by-election circuit even when serious psephologists do not. The information density per contest is just too low for a market to price it.
Why the next-election market barely flinches
Look at the headline UK political contracts and you will not find a Fylde line item. The big liquid markets right now sit on the next general election, the next Labour leader, and the size of the next Reform vote share. Council results filter into those prices only when they form a sustained pattern, and even then with a lag.
Compare that to the way US markets handle their own equivalents. The House control market on Polymarket trades almost daily on the back of generic ballot polls, individual district polling, and special election results, because each US special election is a higher-information event with a bigger media footprint and tighter candidate scrutiny. UK council by-elections, by contrast, are dozens per month, mostly uncontested or uncovered, and the data dribbles in through specialist Twitter accounts and the ConservativeHome digest rather than through any structured feed a market maker can ingest.
The upshot is that micro-elections in the UK are almost a pure inefficiency zone. Not because the results are unimportant, but because the cost of integrating them into a probabilistic view is higher than the marginal information gained from any one of them. If you want to trade UK politics on the back of council results, you are doing the analytical work the market is not doing, and that is either an edge or a trap depending on how disciplined you are about the noise.
The independents question
Worth flagging: both of Thursday's losses came from independents. That matters more than it sounds. The independent vote in English local government has been quietly significant for years, often filling the gap where one of the main parties has effectively given up campaigning. When independents lose seats to a major party, it usually reflects one of two things: a retiring councillor without a clear successor, or a genuine swing in the underlying vote.
The Conservative gain in Fylde reads more like the former on first inspection, a small-coastal ward where the party machine still functions. The Lib Dem gain in Malvern Hills fits a longer pattern of the party hoovering up rural and semi-rural wards from independents and Tories alike. Neither is, on its own, a story. Both, alongside the next forty contests, might be.
The editorial take
Two seats. Not a tremor in the national picture, not a tremor in any market priced off the national picture. But the council by-election calendar is where the underlying tectonics show up first, months before YouGov picks them up and years before a market on the next election starts to move. iPredicta tracks the UK political contracts where these signals eventually surface, on Polymarket and the regulated venues, and the council results worth watching are the ones that fit a pattern, not the ones that make a headline.
Frequently asked questions
Do prediction markets price individual UK council by-elections?
Almost never. The contests are too small, too numerous, and too dependent on local factors for a market to be liquid or informative. Markets that touch UK politics tend to price the next general election, party leadership, or vote share, and council results feed into those prices only when they form a sustained directional pattern across many contests.
Are council by-elections actually useful as a signal then?
Yes, but only in aggregate and only with patience. Analysts who track net seat changes across a year do tend to spot the broad direction of travel before national polling confirms it. The catch is that any single result, like the Fylde or Malvern Hills gains this week, is statistical noise on its own and should not move anyone's view of the next election.
