Fixed industry momentum ranks gave no clear forecast gain across 66 US monthly outcomes in a 2015–2025 archive test
We tested fixed industry momentum ranks with seven dated US source files. Each rank used eleven months of returns and omitted the last source month. We held ranks for up to twelve months after a feasible entry date. The 66 target months from December 2015 to August 2025 did not overlap. Mean monthly rank correlation was 0.0137, with a descriptive 95% capture-year bootstrap interval from -0.0475 to +0.0725. Two simple ranking baselines gave no clear gain. A paper portfolio of the top ten industries gained 0.0194% per month over equal weight in all thirty after an assumed cost of 10 basis points per dollar traded. Its interval was -0.2974% to +0.2928%. A secondary early-minus-later allocation contrast was positive, but the rank contrast remained uncertain. No reliable trading gain was established. Only six capture-year groups support the intervals. The archive is sparse, outcome labels can change, and the paper baskets omit internal stock trading costs. We preserve all ranks, weights, scores, cash ledgers, uncertainty draws, source hashes, code and checks. Related studies use some of the same market outcomes, so this test is not independent confirmation.
industry momentum · archived data · relative return forecasts · chronological test · paper trading · negative result
References
6- usesdoi:10.1111/0022-1082.00146
Established industry momentum motivates the fixed rank rule. This is not a reproduction of the original rule.
- doi:10.1016/j.jfineco.2015.12.002
Prior work studies momentum crashes.
- usesNew months changed six-month industry momentum weights more than revisions on average in ten archived U.S. releases from 2015 to 2026 sci:sha256:4dad13837ba88e4c874fba1e2d3c1f77cc8b00b94b62ccaed073168a22a0c822
We reuse its retained 2026 archive files as possible outcome labels and extend its single-month momentum audit to fixed ranks over longer holding windows.
- usesWith seven archived releases, calendar means raised US industry return forecast error by 5.8 percent in 2016–2026 sci:sha256:0cd06a5f4ba6ad419e7d5245a36ab621fa2c437c0c6023fcf5ea8a2c9f384b4e
We reuse the two retained archived industry releases described by this work and compare a different fixed rank forecast.
- Archived seasonal industry forecasts lost 10.9% a year after assumed costs across five 12-month tests sci:sha256:7bbcffafbafe43442f187e368080634a266911035e2969419e6660429c5b55ef
Related test of fixed seasonal industry forecasts with some of the same source files and target outcomes.
- usesFive published US portfolio return signals did not pass a Holm-corrected holdout test after their original samples to August 2026, and industry seasonality changed to a net loss sci:sha256:6e3392a435bad5ed24d8ecb7f11468605a41e192a217e8dfc6393ca5859be052
We reuse retained source files first described by this study. Its published return-signal tests give prior context.
Cited by
3 publicationsPublic catalog entries on Sciscape.
- usesA 50% mix with common correlations cut joint loss but gave no clear paper risk gain for 30 US industries in 66 months, 2015–2025
We reuse its retained source manifest and feasible entry dates. We test covariance forecasts and constrained paper allocations with new analysis code.
- usesA 50% mix with mean industry risk gave no clear forecast gain in 66 US months from dated 2015–2025 archives
We reuse its source manifest, source loader, fetch script and seven dated target windows. The new test changes marginal risk estimates.
- Shrinking ten-year industry means gave no clear relative-return forecast gain in 71 monthly archive tests, 2016–2026
This prior test uses industry momentum and some of the same source periods. It is not independent evidence for the present result.