This is called burying. It makes sense in systems that violate the later-no-help or later-no-harm criteria, but instant-runoff voting satisfies both of those.
https://electowiki.org/wiki/Tactical_voting#Burying
I like this framing a lot – thanks for sharing!
Looking forward to it! Will it be on Pocketcasts?
Is ED different to the Coombs method?
https://en.wikipedia.org/wiki/Coombs%27_method
Can you explain why "most democratic Western countries will become fascist dictatorships" specifically is the most likely outcome? I'm not quite following the "line go up" reasoning.
Just to give some concrete illustrations: Western Europe has dropped by 0.08 points on the Democracy Index over the last decade. If that happens again, it will fall to 8.28, so about as democratic as the UK is now. Definitely not fascist dictatorship territory!
If we pessimistically assume that Western Europe will decline faster in line with the world average, we get to 8.13, a bit more democratic than France. Also nothing close to a fascist dictatorship.
If Western Europe declines twice as fast as the world average, we get 7.9, a bit more democratic than the US. Three times as fast, we end up just above Belgium. Double that, and we'd be about level with Poland and India: not functioning democracies, but still democratic enough that upsets can happen (e.g. Modi's BJP just lost a major state election in Karnataka). I certainly wouldn't call either of those countries fascist.
Let's assume the decline in democracy speeds up ten times over the next decade: then Western Europe still looks more democratic than Mexico, Ukraine and Peru. A sudden tenfold speed-up in current trends sounds like a tail risk to me, but it doesn't come close to the outcome you're predicting. How do you get to a prediction as strong as that one given the data you've presented?