Methodological audit of power-law claims in supply chain operational data: analysis code, fitted parameters, and citation list

This deposit contains the analysis code, fitted parameters, output files, audit records, and citation list supporting the manuscript "Auditing power-law and scale-free claims with uniform Clauset-Shalizi-Newman re-analysis: evidence from supply chain operational data" by Bruce Rishel (Independent Scholar, ORCID 0009-0000-8324-4331).

The audit applies the Clauset-Shalizi-Newman 2009 maximum-likelihood estimation protocol with semi-parametric bootstrap goodness-of-fit testing and likelihood-ratio tests against alternative distributions, plus the Broido-Clauset 2019 five-category corpus classification, to a verified population of 20 published power-law claims in the supply chain operational literature. Three foundational datasets are re-analyzed across five constituent distributions: Acemoglu et al. 2012's sectoral input-output network (BEA 1997 Use Table), Liu, Shen, Tan 2021's international trade product networks (BACI HS92 Y2024 current release, 1,217 networks), and Axtell 2001's US firm-size distribution (SUSB 2017, n = 5,996,900 firms via the Virkar-Clauset 2014 binned-data MLE). The remaining 17 audit targets are characterized at methodology rigor level without independent data re-analysis. Every audit-target citation is verified by direct publisher-record check under a quintuple protocol (title, authors, DOI, claim text, and author-list completeness).

Contents of the deposit:

analysis/: Python implementation of the audit's re-analysis pipeline: BACI parsing and corpus fitting, BEA 1997 Use Table reconstruction, SUSB binned-data MLE via Virkar-Clauset 2014, Vuong likelihood-ratio tests, Broido-Clauset 2019 corpus classification, semi-parametric bootstrap goodness-of-fit, and cross-validation of the Acemoglu sectoral-network re-analysis (d_i and v_i) against the published Alstott et al. 2014 powerlaw Python library (fitted exponents and Kolmogorov-Smirnov distances match to four decimal places).

outputs/: JSON output files containing per-target fitted parameters, x_min values, KS distances, bootstrap p-values, Vuong test statistics, and corpus aggregate statistics; powerlaw-library cross-validation output.

audit_records/: Per-target audit records for the three re-analyzed datasets (Acemoglu et al. 2012, Liu-Shen-Tan 2021, Axtell 2001), including methodology grade, data acquisition tier, sample sizes, fit details, and audit verdict.

audit_targets/: Structured audit-target list of all 20 verified entries with author, title, venue, year, DOI, verification status, data tier, methodology rigor level, and re-analysis results.

verification_protocol.md: The quintuple citation-verification protocol and six-category defect taxonomy.

references.md: The manuscript's numbered reference list.

figures/: Three figures from the manuscript: BACI corpus alpha distributions, Acemoglu d_i CCDF with fitted power law, SUSB firm-size CCDF with fitted power law.

README.md: File inventory, dependency versions, reproducibility instructions, and citation guidance.

CHANGELOG.md: Version history of this deposit.

LICENSE: MIT License (code), CC-BY 4.0 (documentation).

Data sources used by the audit code are publicly available at no cost: BEA 1997 detailed Use Table (Bureau of Economic Analysis); BACI HS92 Y2024 release (CEPII); SUSB 2017 US enterprise-size brackets (US Census Bureau). Source URLs are documented in the README.

Reproducibility: the audit's central verdicts (three of five re-analyzed distributions reach plausible status under bootstrap goodness-of-fit testing; fitted exponents differ from originally reported values) can be reproduced end-to-end by following the README's instructions on a standard Python 3.11+ environment with the listed dependencies (numpy, scipy, openpyxl, powerlaw, matplotlib, pandas).

Citation: Rishel, B. (2026). Methodological audit of power-law claims in supply chain operational data: analysis code, fitted parameters, and citation list. Zenodo. https://doi.org/10.5281/zenodo.20335206