Andersson's consensus odds fell to 77.6% overnight into 2 September, the lowest reading we have recorded

A single overnight jump

At 23:29 UTC on 1 September 2026 the consensus figure gave Magdalena Andersson 86.1% to become prime minister after the election, and Ulf Kristersson 12.3%. The next observation in the series does not arrive until 09:49 UTC the following morning, and by then the figures stood at 77.6% and 20.8%. Those are our two closest data points bracketing the move, there is no reading between them, so this is a single jump between two fetch cycles roughly ten hours apart, not a stepped move over a few hours.

Measured against exactly 24 hours before the latest observation, by our own method for the 24-hour figure shown on the market page, the fall is 8.5 percentage points for Andersson.

Both venues moved by about the same amount

At 09:49 UTC Polymarket priced Andersson at 77.6% and Kristersson at 20.9%. Kalshi priced the same pair at 78.4% and 17.7%. Both venues thus moved by about the same amount at the same time, Andersson down a little over eight percentage points on Polymarket and seven on Kalshi, which is the single most important check on a move of this size: a figure that moves by about the same amount on both platforms is a fact about what the markets as a whole are now pricing, not noise from one order book.

A new low, not a return

77.6% is the lowest consensus figure we have recorded for Andersson in this market's history. The series' highest reading was 92.1%, set on 23 August 2026. The market's very first observation, on 3 August 2026, gave her 80.5%, meaning this morning's figure sits below even the series' opening reading, not back at some earlier level.

For comparison: at 14:04 UTC on 14 August 2026 consensus stood at 81.9% for Andersson. This morning's figure sits lower than that too, the market has not returned there, it has passed through that level on its way down.

What we are not saying

We do not know why the market moved, and we are not going to guess. We publish what the price series shows and the sources it came from: the current consensus figure gives Andersson 77.6%, it does not give her an explanation. A claim about cause needs a source a reader can check, and we have none tying this move to a poll, a televised interview or an event.

That is not an apology, it is the point: if the cause emerges later, and eleven days out from an election it often does, the series stands exactly as it was, timestamped and sourced, and can be read against whatever comes to light. A price chart with an explanation attached after the fact is a worse historical source than one without.