At a Glance
Goods audited that had an independently evidenced claim on either the Planet or People & Supply Chain dimension — from any of three review sources.
Candidate entities narrowed to one legal entity via commercial supply-chain retrieval — testing whether specialist data closes the gap public sources leave.
Dimensions admissible even from paid, compliance-grade intelligence — confirming the deficit is structural, not a sourcing failure.
Abstract
The United Nations Guiding Principles on Business and Human Rights establish a corporate responsibility to prevent and account for adverse human rights impacts across value chains, including those arising through environmental harm. Fifteen years on, the Office of the High Commissioner for Human Rights notes that accountability gaps persist.
Such gaps are addressed through several mechanisms; this study examines a precondition of one: demand-side pressure at the point of purchase. Labor rights violations cut production costs; certification without independent verification lets that saving hide behind a claim; and the premium paid for the certified good flows to the concealing firm rather than to remedy. Such pressure depends on evidence reaching the terminal node of the chain, and research makes that evidence necessary but not sufficient: disclosing concealment reduces intent toward the concealing actor, while price and availability still dominate what people buy. The study therefore asks not whether disclosure changes conduct, but whether the evidence it depends on is available at all.
Two propositions are tested: that the gap is one of distribution rather than absence, and that no single source class covers every dimension due diligence requires. Both hold. An audit of six goods found no assessor covering more than three of five dimensions, and no evidenced claim on the environmental or labor dimension for any good. A retrieval against commercial supply-chain intelligence was admissible for two dimensions and rejected for three: the inputs are largely public, but the entity resolution needed to use them is not.
The precondition is not met. A third proposition follows without separate test: any composite blends dimensions that are backed by entirely different bodies of evidence. The study specifies a non-compensatory architecture accordingly, and a rule for disclosing compliance-grade evidence to non-expert readers without implying harm it does not establish. Whether meeting the precondition delivers remedy is not tested here.
Keywords
Central Claims
The evidence needed for good-level due diligence is unevenly distributed, not missing. Its raw inputs are largely public. What is proprietary is the entity resolution that makes them usable.
No single source class covers all five dimensions material to due diligence at the level of the individual traded good.
Each dimension rests on different sources of different quality. A composite rolls them into one number and hides the unevenness.
Derived from P2 — not tested separatelyContents
Appendices A–D — reproducibility statement, Sayari retrieval parameters, the downstream Value prong, and the assisted-research protocol. 42 footnotes with full source citations.
Cite This Work
Du, A. (2026). Human Rights Due Diligence, Information Asymmetry, and the Limits of Good-Level Transparency in Global Value Chains. NYU M.S. Global Affairs Capstone, Center for Global Affairs. https://github.com/alx-du/product-integrity-scorer/blob/main/CAPSTONE.md