Mom36m

Long-Run Reversal — De Bondt and Thaler 1985

Reversal

Quick facts

Factor ID Mom36m
Display name Long-Run Reversal
Family Predictor
Category Reversal
Direction lower_expected_return
Status supported
Paper De Bondt and Thaler 1985
Source tables daily
Required fields ret
PIT required no
Coverage 2013-02-28 – 2026-05-15 · 483,827 rows · 70.6% of panel

Formula

prod(1 + ret over t-504..t-253) - 1

Compounded return between roughly 36 and 13 months ago; long-run winners tend to reverse over the next year.

China adaptation

Trading-day windows are used so that suspensions and holidays do not silently shorten the formation period.

Direction-adjusted cumulative return

0.99 1.79 2.59 3.38 4.18 2013-02 2016-05 2019-09 2023-01 2026-04

Return metrics

Metric Value
Months in sample 159
Monthly mean 0.9304%
Monthly std 3.0257%
Annualized Sharpe 1.07
t-stat (simple) 3.88
Hit rate (months > 0) 66.7%
Last 12M compounded 9.22%
Cumulative wealth (start = 1) 4.059

Recent 12 months

Month L/S return Cumulative Names
2025-05 +3.22% 3.836 4526
2025-06 +2.06% 3.915 4562
2025-07 -1.75% 3.846 4579
2025-08 +1.81% 3.916 4606
2025-09 +0.29% 3.927 4649
2025-10 +2.10% 4.010 4677
2025-11 -2.43% 3.912 4714
2025-12 +2.71% 4.018 4745
2026-01 +1.39% 4.074 4761
2026-02 +2.67% 4.183 4775
2026-03 -0.89% 4.146 4786
2026-04 -2.10% 4.059 4792

Known data issues

Pre-2012 panel coverage is thinner so the early sample has fewer eligible names.

Notes

A-share long-run reversal is sensitive to delistings and major regime changes such as the 2015 bubble.

See also: Reversal family, Factor library, Factor returns.