Ergunsu-Yang. Why the models missed the surge

Replication package for "Why the models missed the surge: decomposing the forecast failure in China's electric heavy-truck transition, 2021–2026." Contains all analysis code (Python, pinned environment), the assembled monthly data panels with registry-basis and provenance tags (new-energy and LNG heavy-truck registrations, chained diesel and LNG price series, IEA GEVO 2026 anchors), a consolidated parameter-extraction audit record for the forecasting baselines, a primary-source verification record with URLs, and reference outputs with SHA256 checksums.

Running run_all.py reproduces every table and figure in the manuscript deterministically (fixed seed), including the calibrated fleet cost-sensitivity (lambda = 1.66 per CNY/km) and the 32-run Shapley decomposition of the 2025 forecast wedge. All underlying statistics are from public sources cited in the documentation.