Work

Sector Correlation and Subsequent Drawdowns: A Preregistered Chronological Evaluation

Year: 2026
Type: article
Abstract: Neither primary hypothesis met its preregistered support conditions. This study evaluates whether a twenty-session mean correlation among nine long-history US sector ETFs is associated with subsequent closing-price drawdowns. Data end on 4 November 2026; the chronological evaluation begins in 2015. All nineteen registered tests and all cost sensitivities are retained. The study is retrospective and does not establish a prospective alarm or timing skill. The VIX-adjusted statistic and restricted linear VIX forecast comparison have calibration limitations. The paper reports these and the hypothetical cash-proxy and event-concentration limitations. Exact reproduction requires lawful access to the recorded inputs; raw commercial prices and internal daily derived series are not redistributed. The publication copies of the protocol normalize one historical implementation path. The original and presentation hashes are both documented, and the frozen original remains unchanged. The comparison of 253 readings, maximum absolute difference 2.22 × 10⁻¹⁶, tolerance 10⁻¹², and nineteen-return negative control difference 0.08699 are retained. Generative AI tools assisted in preparing this work; the author is responsible for all content. Author: Peilun Cai, ORCID 0009-0007-5364-6883. Manuscript and original companion materials: CC BY 4.0; third-party material retains its own terms. Independent review of these publication-only changes is pending.
Author: Cai, Peilun
Language: English
Subfield: Finance
Open Access status: green
OpenAlex ID: W7233556630
FWCI:
Cites:
Cited by:
Related to: