Scott Alexander at Astral Codex Ten: Some of the party’s problems are hard and have no shortcuts. But the big one – figuring out whether replacing Biden would even help the Democrats’ electoral chances – is a good match for prediction markets. Set up markets to find the probability of Democrats winning they nominate Biden, vs. the probability of Democrats winning if they replace him with someone else.
(see my Prediction Market FAQ for why I think they are good for cases like these)
Before we go into specifics, the summary result: Replacing Biden with Harris is neutral to slightly positive; replacing Biden with Newsom or a generic Democrat increases their odds of winning by 10 – 15 percentage points. There are some potential technical objections to this claim, but they mostly suggest reasons why the markets might overestimate Biden’s chances rather than underestimate them.
The rest of this post is mostly sources and technical discussion trying to establish that this is true. You can skip it if you don’t care. But in case you do:
John Stossel and Maxim Lott have a site called Election Betting Odds which tries to aggregate information from many different betting markets to answer questions like these. Here are their claimed headline results – this is probability conditional on being nominated:
