Before and after the fireworks last night, I was glued to my television watching BBC coverage of the 2024 UK election. The final results were very consistent with expectations: a comfortable majority for Labour even though they didn’t really expand their vote share, as they exploited shifts away from the SNP in Scotland and away from the Conservatives to the new Reform UK party in England. The Conservatives are down but not, unlike the 1993 Progressive Conservative party in Canada, out; here Reform finished 3rd in the popular vote but only managed to gain 4 seats, the same as the Greens and Plaid Cymru. Meanwhile the centrist Liberal Democrats, despite finishing 4th in the popular vote, grew their seat count to an all-time high of 71.
One has to admire British efficiency when it comes to running an election. Polls closed everywhere at 10pm. Paper ballots only. No early voting. Mail-in voting is allowed, but the vote has to be received by 10pm for it to count. This was the first election for which voter photo ID was mandatory, which was mildly controversial but does not appear to have depressed turnout much if at all. Counting starts at 10pm, and – here’s the craziest part – no preliminary results are released! As such, for the first hour-and-a-half of TV coverage, there was nothing to talk about except the exit polling, which has historically proven to be exceptionally accurate. The swiftest seats reported before midnight, and most seats between 3am and 6am. By shortly after 5am, enough winners had been officially declared to ensure a Labour majority government; a few hours after that, the transition of power was formally completed and Keir Starmer became prime minister.
Of course, the US has five times the population of the UK and two-thirds the number of constituencies, so the vote to be counted in a typical US congressional district would be something like 7.5x that of a typical UK riding. Plus, in the UK there was only the one election on the ballot, because local elections are held at a completely separate time. Still, how nice it would be to have complete finality about vote-counting within hours of the polls being closed.
On PredictIt, the Biden contract had recovered from the steep dip on July 3rd that prompted my last post, closing the day back at $0.47. Since then things have again moved a little against Biden; at this writing the Biden contract is at $0.40 while the Harris contract is at $0.48. Today Biden is giving a lengthy interview to ABC, which will be aired on a tape-delayed basis tonight. Silver’s model has ticked up to 71% for Trump over Biden based just on polling updates (i.e., the model doesn’t know about the debate performance and the emerging controversy over Biden), and now sees Trump as the popular vote favorite also.