Replication package for Middle East escalation and systemic food-inflation risk analysis

This archive contains the replication package developed by Sreejith Aravindakshan for the analysis of Middle East escalation and systemic food-inflation risk across 39 countries. The country sample includes economies classified as low-income, lower-middle-income, and upper-middle-income, and excludes high-income countries. The package provides the R workflow, shareable input datasets, metadata, codebooks, and documentation required to reproduce the manuscript’s fixed-effects model results, scenario projections, GIS-ready outputs, figures, tables, logs, and quality-assurance checks. The archive includes: run_released_workflow.R: the main workflow runner. R/: stepwise R scripts for setup, input validation, shock-index construction, panel preparation, fixed-effects estimation, scenario projection, GIS-layer preparation, figure-data preparation, individual figure generation, combined figure generation, table export, and warning checks. R/functions/: helper functions, plotting themes, palettes, QA utilities, read/write utilities, and upstream workflow components. inputs/analysis_ready/: the analysis-ready country-year panel used in the fixed-effects models. inputs/food_cpi/: the curated 2025 food-inflation baseline used for the corrected Extreme-scenario GIS projections. inputs/metadata/: country-level metadata, including population and multidimensional poverty indicators used for exposure and welfare-burden weighting. inputs/shock/: historical and 2026 YTD Middle East shock-index inputs, accepted benchmark files, component data, and shared external shock-related indicators. inputs/scenario/: scenario input files used for bounded stress-test projections. docs/codebooks/: variable-source documentation and codebook materials. docs/methods/: reproducibility and data-source notes. docs/manifests/: workflow manifest files documenting expected outputs and package structure. The workflow is designed to be run after unzipping the archive by opening a fresh R or RStudio session and sourcing the main runner script: source("run_released_workflow.R") The runner automatically sets the project root to the ReleasedCode folder, so the workflow does not depend on the original development path. Generated outputs are written to outputs/, including model outputs, QA files, GIS-ready datasets, figure inputs, individual and combined figures, tables, and run logs. The main analytical purpose of the package is to reproduce a 39-country workflow that links a constructed Middle East geopolitical shock index to Food CPI growth using fixed-effects models, applies an escalation-specific positive-shock coefficient for bounded Extreme-scenario stress testing, and maps projected food-inflation changes relative to actual 2025 food-inflation baselines. Diesel-price and urea-input interaction models are included as mechanism evidence, while the headline GIS projection uses the asymmetric positive-shock model to avoid misleading country-specific interaction offsets. The package is intended for research transparency, reproducibility, peer review, and public archiving. It includes shareable processed and benchmark-controlled inputs required to reproduce the released outputs, but does not include restricted raw source files, local development folders, or interactive R history files. Users should run the workflow in a clean R session with current tidyverse-compatible packages. In particular, rlang version 1.1.7 or newer is recommended, as older versions may cause workflow errors on some systems.