The Metaculus Democracy Threat Index
Read the original on Astral Codex Ten →
Summary
Analyzes Metaculus's new Democracy Threat Index (153 forecasting questions on US democracy) as a forecasting-based alternative to expert-driven democracy indices, prompted by Scott's own 'am I suffering Trump Derangement Syndrome?' self-check. Advantages: transparent, crowdsourced (limits ideologue fingers-on-scale), forecasting-based fine-grained probabilities (10% vs 50% chance of cancelled elections, not just binary). Remaining risks: crowd-attack (solved by moving to real prediction markets), and question-SELECTOR bias (the memecoin-corruption vector as a uniquely-Republican-coded inclusion -- are they allowed to judge one side has more?). Key insight: because a stable question set locks in 2025's biases, the index becomes MORE trustworthy over time as biases drift; plus the reassuring finding that forecasters expect democracy not to worsen over Trump's term, and the Dec-2025 threat-level decline.
Why this score
Quality 69 · Strong. Strong: a thoughtful, useful analysis with a genuinely fresh point (a stable question set gains trustworthiness as biases drift) and the sharp memecoin-vector bias example, but shorter and topical.
Claude’s paradigm shift 50 · Moderate. Notable: the 'biases-drift-so-a-frozen-question-set-becomes-trustless-over-time' argument and the conditional-market democracy-index idea are non-obvious contributions.
Real-world impact 2 · Minor. Minor/within-discourse: an analysis of a forecasting index; no material footprint.