Millions Vote as Starmer Faces His Worst Night Yet

Prediction markets say local election losses won't end the prime minister's tenure.

Polling stations opened across England and Wales on Thursday for local elections that are expected to deliver Labour its sharpest rebuke since taking office. Council seats in metropolitan strongholds are in play, Reform UK is polling at historic highs, and the question dominating Westminster is not whether Labour will lose ground but how much.

The academic and media consensus has hardened fast. The London School of Economics published a piece on Thursday titled "The end of the Starmer government." The New York Times framed the day as Starmer bracing for losses "amid a new era of UK politics." The narrative is one of terminal decline: a prime minister battered by a 48-year low in consumer confidence, sticky inflation fed by the Iran war energy premium, and a civil service row that alienated his own Whitehall machinery.

The money tells a different story. On Polymarket, the probability of Starmer leaving office by May 15 sits at 12% as of May 7, 2026. That figure was 34% in April, when the civil servant dismissals and the consumer confidence nadir hit simultaneously. The collapse in that probability over the past month is the market's verdict: local election losses are priced in. They are not priced as fatal.

The gap matters. Political commentary trades in momentum and mood. Prediction markets trade in money. And right now, traders with nearly $89,000 in 24-hour volume on the contract are overwhelmingly betting Starmer survives the week. A bad night for Labour, losing councils and possibly flagship boroughs, does not automatically become a leadership crisis. Local elections in the UK rarely do.

What to watch: the speed and scale of Labour losses as results arrive Friday morning, whether any cabinet figures break ranks publicly, and whether Reform UK translates polling strength into actual council seats. If Labour's losses stay within the expected range, Starmer walks on. If they dramatically overshoot, that 12% starts moving.