Does Cross-Sectional Dispersion actually work?

DEAD

No edge left once tested on a fair footing.

The claim

Low cross-sectional return dispersion (stocks moving together calmly) signals risk-on — Dispersion / correlation tradition

What we did

Tested against Buy & hold SPY (gates) / same book no exit (overlay), 2016-2026 (survivorship-free panel), overlay on SPY.

The result

Result: 0.04 dSharpe (vs buy-hold)

dSharpe -0.25 (IS -0.80 -> OOS +0.04); fails the data-mining floor as a standalone equity gate

Verdict

As a simple low-dispersion = risk-on equity gate, cross-sectional return dispersion is noise (below the noise floor). NOTE: this is ONE construction -- IMPLIED dispersion and realized-correlation regime are different signals with different uses (vol trading), and are not tested here. So this is 'return-dispersion-as-equity-gate fails,' not 'dispersion fails.'

What the test did and did not do

tested out of sample
point-in-time data
checked for look-ahead
net of real costs

Almost nobody publishing a backtest says which of these they did. We list the ones we failed as well as the ones we passed.

Tested 2026-06-23. The same treatment, applied to every other system, is on the test bench.

← CAPE / P-E10 (as a timing signal) All findings Curve 10Y2Y>0 →

Written with AI assistance from our own test database. The numbers are computed, not composed.
General market commentary for education, not personalised investment advice. Backtested results are hypothetical. Trading carries a real risk of loss.