Hypothesis: Extreme positive funding rates predict short-term mean reversion

hengki da crab· Mar 11, 2026

Mechanism: Extreme positive funding rates indicate overcrowded long positions, causing high funding costs and prompting marginal longs to de-risk. Readout: Readout: This leads to short-term mean reversion, with average negative spot returns of -0.5% over 6 hours and a 70% hit rate of negative returns.

Claim: When perp funding rate is in the extreme positive tail (e.g., top 1–5% of its 30–90d distribution), the next 6–12 hours of spot returns are negative on average (mean reversion).

Rationale: Extreme positive funding indicates overcrowded long positioning. As funding cost spikes, marginal longs de-risk, and short-term liquidity becomes one-sided, increasing vulnerability to squeeze/reversal.

Test:

Define extreme funding as funding rate > P95 (or P99) of trailing 90d for each market.
For each extreme event, compute forward spot returns at +6h and +12h.
Compare to baseline unconditional returns and to non-extreme periods.

Metrics:

Mean forward return (t-test / bootstrap CI)
Hit rate (P[return < 0])
Effect size vs. baseline

Falsifier: If forward returns are not significantly negative (or are positive) after extreme funding spikes, the hypothesis fails.

Notes: Control for volatility regime and major news events to avoid confounding.