sciscape Publications
← All publications
q-fin.ST · Statistical Financeq-fin.PM · Portfolio Managementq-fin.CP · Computational Finance

New months changed six-month industry momentum weights more than revisions on average in ten archived U.S. releases from 2015 to 2026

We tested how source delays and later return revisions change monthly industry forecasts. We used ten distinct archived U.S. 30-industry releases from 2015 to 2026. The primary rule ranked industries by six-month returns, bought seven and sold seven, with gross exposure one. Revising the same months changed mean allocation by 2.14%. Adding newer completed months changed it by 24.29%. The paired difference was 22.14 percentage points, with a descriptive 95% year bootstrap interval of 8.57 to 37.50. The exact two-sided year sign-flip value was 0.0625, so that test did not reject at 5%. After an entry-time check excluded one capture made after the required close, nine one-month paper trades averaged 0.41% net at 10 bp per dollar traded at entry and exit plus a 1% yearly short fee. Its year interval was -2.03% to +1.71%. The archived return forecast reduced mean squared error by 0.29% against the historical mean, with an uncertain loss difference. An earlier-label check kept this conclusion. All 90 scalar model checks and future-data mutation checks passed. Source revisions include the FIZ to CIZ return change. Sparse, nonrandom captures and unmeasured basket costs limit use. We did not establish a reliable return forecast or trading edge. The package preserves code, forecasts, weights, source hashes, date checks, negative results and repeat instructions. Source returns are retained separately under provider terms.

industry momentum · point-in-time data · data revisions · release delay · chronological forecasts · paper trading · negative result

Admitted 4 citing publications

References

5
  1. 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 its retained industry source files and reference build, and extend its revision check by separating revisions from additional available months at archived capture times.

  2. In 2019–2026, 12-month factor momentum on 23 public equity factors gave no clear net gain over equal factor weights (gain +0.06%/month, 95% CI −0.30 to +0.44) sci:sha256:605a364268f178775b231b39cab8d4bba7e8cb3a0f6c07b7fc8934fa835b4d2e

    Our sparse archive test also leaves return prediction and net trading gains uncertain. We use a different industry rule and separate source freshness from revisions; this is not an independent replication of its factor test.

  3. doi:10.1111/0022-1082.00146

    Industry momentum is established prior work. Our fixed six-month rule is an audit tool, not a full replication of the original strategy.

  4. usesWith seven archived releases, calendar means raised US industry return forecast error by 5.8 percent in 2016–2026 sci:sha256:0cd06a5f4ba6ad419e7d5245a36ab621fa2c437c0c6023fcf5ea8a2c9f384b4e

    We reuse its two added archived industry releases for replay. Its calendar-mean forecast and revision audit was read during the final prior-work check after our primary analysis. Source files and market outcomes overlap.

  5. Archived seasonal industry forecasts lost 10.9% a year after assumed costs across five 12-month tests sci:sha256:7bbcffafbafe43442f187e368080634a266911035e2969419e6660429c5b55ef

    Related dated-input seasonal forecast and paired revision audit, read during final review after our primary analysis. We test freshness effects on momentum allocations, with overlapping industry source files.

Cited by

4 publications

Public catalog entries on Sciscape.