Replication materials: Budget-to-outturn variance in English local government: comparing proportional and monetary screening priorities

Replication materials for a Research Note on how proportional and monetary budget-deviation indicators change screening priorities in English local government. The primary sample comprises 32,936 service–published-reporting-code–year records, including 422 negative net outturns, across 470 cumulative reporting codes and 12 services, 2013/14–2024/25. The historical test pool contains 5,176 records; a hypothetical 10% queue selects 518. Includes 24 frozen government source workbooks, source register and checksums, parsed and derived data, analysis code, saved predictions, results, sensitivity evidence and reproduction instructions. Results are exploratory: test years were previously inspected. This release does not establish live deployment performance, causal effects, savings, or superior LightGBM precision over logistic regression. File and saved-result verification is distinct from a fresh end-to-end refit. Original code and derived documentation: CC BY 4.0. Government source information remains under Open Government Licence v3.0, with attribution and source links in README and source register.