Replication package for Container shipping disruptions and the economics of decarbonization
Description
This is the complete replication package for the paper "Container shipping disruptions and the economics of decarbonization" (Hugo Morão, 2026). It contains all MATLAB code, input data, intermediate estimation output, and final figures needed to reproduce every result in the paper (Figures 1–32). The study estimates a suite of Bayesian vector autoregressions (BVARs) to trace how container-shipping disruptions propagate to global macroeconomic variables and to climate-finance and decarbonization outcomes. The package is self-contained: it bundles the required third-party toolboxes (BEAR 5.2.2, the Empirical Macro Toolbox BVAR_, and export_fig) and provides cached posterior draws so all figures can be rebuilt in minutes, or re-estimated from scratch. A single driver script (main_replication.m) runs the whole workflow. See README.pdf for full instructions and a figure-by-figure map to the code and data. Licensed under CC BY 4.0.
Files
Steps to reproduce
Software: MATLAB R2021b or newer, with the Statistics and Machine Learning Toolbox and the Optimization Toolbox. All third-party toolboxes needed (BEAR 5.2.2, Empirical Macro Toolbox BVAR_, export_fig) are bundled under toolboxes/ and added to the path automatically, no separate installation is required. 1. Download all files and extract, keeping the folder structure intact (data/, settings/, results/, figures/, toolboxes/, and the scripts at the root). 2. Open MATLAB and change the current folder to the extracted package folder (the one containing main_replication.m). 3. Run the driver as a whole script (Run button, or type main_replication in the Command Window — do not run it section-by-section): main_replication It self-configures all paths and runs check_requirements to confirm your environment. 4. Fast reproduction (default, ~minutes). With replication_mode = false, the script loads the provided posterior draws in results/ and writes figures/Figure1.png … Figure32.png. These are exactly Figures 1–32 of the paper. 5. Full reproduction (~hours). Set replication_mode = true near the top of main_replication.m and rerun. This re-estimates all 31 Bayesian VAR specifications with the BEAR toolbox from the raw data in data/, overwrites results/, then rebuilds the figures. A fixed seed (rng default) is set for reproducibility. Output mapping: figures/FigureN.png is the paper's Figure N. The file figure_map.csv (and Section 6 of README.md) lists, for every figure, the code line that builds it, the specification/settings file, the results/*.mat it reads, and the input data file. Integrity check (optional): the package ships environment.lock.json (MATLAB environment plus SHA-256 of every input and output). Run verify_lock in MATLAB to confirm no data, settings, or results file has been altered.