Replication package for "Financed but Unstudied: Instrument Labelling and the Measurement Blind Spot in Green Finance for Resource-Intensive Value Chains"

Data and code reproducing every result in a study that pairs three independently assembled datasets: a PRISMA 2020 systematic mapping of the green-finance literature on resource-intensive industries in emerging economies (4,423 records identified, 185 retained); an original, source-verified database of 41 development-finance-institution transactions in the global cotton-textile value chain (2015–2026), each with a primary-source URL; and a third Lens.org search designed to test the paper's proposed mechanism directly by pairing the same sector terms with financing vocabulary that excludes every labelled-instrument term. Contents: raw Lens.org exports; the complete screening pipeline; a record-level screening log giving the decision and reason for all 4,423 identified records; the final 185-record corpus; the transaction database; the instrument-by-sector gap matrix; a ten-specification robustness analysis; and the record-level substantive classification of all 30 title-matched Search R records. The analysis scripts assert every computed statistic against the value printed in the manuscript and exit non-zero on any divergence, so a successful run of run_all.sh is itself a verification of the reported figures. Screening and classification were performed by a single coder with disclosed AI assistance for mechanical text processing and for a substantive verification pass. There was no second human coder and no inter-rater reliability statistic is reported.